WikiLean · Articles Articles Brain Contribute About

Preliminary: these links have not yet had an independent audit.

Gradient

98 passages48 formalized18 partial32 not formalized

Statement only

More ways to filter
Library
About this page

This page

  • Wikipedia revision 1355918370, pinned: the text here is that revision, not today's article.
  • WikiProject Mathematics: C-class, High importance
  • Annotated article on WikiLean.
  • 98 passages checked, 108 Lean locations (and 47 related declarations).

Central statements: 56 of the 98 passages state the article's subject or a named result or definition; the other 42 are supporting (examples, setup). Of these 56: 29 formalized, 9 partial, 18 not formalized.

What the colours mean

  • Formalized a Lean declaration states it, proved, or a Lean definition defines it
  • Statement only Lean states it, but the proof is not finished
  • Partial Lean states part of it, or a more general or special form
  • Not formalized searched every library and not found

Libraries searched

Searched 48 Lean libraries, pinned 2026-10-17, 3 left out (library set u1).

How these links were made

  • Accuracy, from an audit of a sample of 224 items (blind, not independent): Strong links (exact, close or broad match) on supporting passages that have exactly the right relation: 92% (95% interval 49–99%), from 36 drawn of 1029, weighted by relation and library, an effective sample of 5.1; Links in a counting relation (exact, close, broad, narrow match or part) on supporting passages that have exactly the right relation: 73% (95% interval 40–92%), from 48 drawn of 1740, weighted by relation and library, an effective sample of 8.1; Central and subject passages whose formalization status is right: 86% (95% interval 73–94%), from 44 drawn of 1905, weighted by recorded status, an effective sample of 42; 'Not formalized' on supporting passages that is right (no declaration in the 48 libraries states it): 60% (95% interval 25–87%), from 12 drawn of 1236, weighted by article, an effective sample of 6.1.
  • Audit run tierA-batch1-audit, run by jack, 2 Nov 2026, 21:10 – 3 Nov 2026, 02:10 UTC: judged by Claude Opus 5.5 (claude-opus-5-5).
  • Record set tierA-b1b (cf-locate-recordset/2, statuses by status rule 3), exported 1 Nov 2026 by formal-companion at commit d85385a.
  • Seal sha256 23710337d053e8212378a9e580805628bcdfda61a6fc431e65f24a4a16c9e602.
  • Inventory run inv-tierA-b1r, run by jack, 30 Oct 2026, 07:23–20:00 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5). It listed the formal passages, each with its kind and prominence, that the locate run then searched.
  • Locate run loc-tierA-b1b, run by jack, 31 Oct 2026, 04:59 – 1 Nov 2026, 03:06 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5).
  • Challenge run chal-tierA-b1, run by jack, 1 Nov 2026, 03:19–14:25 UTC: worker Claude Opus 5 (claude-opus-5), reviewer Claude Opus 5 (claude-opus-5). In this article it re-checked every Lean location and absence on a central or subject passage (69 locations and 18 absences); on supporting passages, a seeded sample (seed 0; drawn at 0.5 for strong matches, 0.2 for weaker links and 0.25 for absences): 24 of 39 locations and 3 of 13 absences. In all, 93 of 108 Lean locations (86%) and 21 of 31 absences (68%) were re-checked.
  • Passage text: 98 of 98 passages shown in full, each the slice of the pinned revision's text that its record's span sha256 hashes, checked when the store was built.
  • Lean code: each declaration's type, from the declaration corpus of universe u1 (built 18 Oct 2026, sha256 a364fd6aede5), at the commit its card links.

How the text is shown

Each passage is shown in full, as revision 1355918370 states it (checked against its record's span hash when the store was built). Where a passage is only part of a sentence, it is highlighted and the rest of the sentence is shown in grey around it, from the same copy of the revision; a list item also shows the sentence that introduces it. Formulas are drawn from their TeX; hover one to see it.

Each card's Lean code is its declaration's type as Lean prints it, from the declaration corpus of library set u1 at the card's commit, laid out as a header: binders before the colon, then the hypotheses as arrows (the corpus keeps no hypothesis names) and the conclusion, in Mathlib's notation (𝓝, π, R[X], Type*). It is not the source text; Source opens that.

Passages

What “exact match”, “more general” and the other labels mean
  1. Definition · Subject of the article

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    In vector calculus, the gradient of a scalar-valued differentiable function f of several variables is the vector field (or vector-valued function) ∇f whose value at a point p gives the direction and the rate of fastest increase.

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib defines the gradient for scalar functions on any complete inner product space over a real or complex scalar field, through the duality between the space and its dual.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d14af5aaa05d · pair key 3fa0b1007b94.
  2. Theorem · main statement

    Not formalized
    The gradient transforms like a vector under change of basis of the space of variables of f.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 49596ad84f93).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  3. Theorem · main statement

    Partial

    Partial: TauCeti states part of this (proved).

    If the gradient of a function is non-zero at a point p, the direction of the gradient is the direction in which the function increases most quickly from p, and the magnitude of the gradient is the rate of increase in that direction, the greatest absolute directional derivative.[1]

    Show in article · Read it in context on Wikipedia

    TauCeti.norm_gradient_eq_norm_fderiv states part

    theorem TauCeti.norm_gradient_eq_norm_fderiv {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] (f : F → 𝕜) (x : F) : ‖gradient f x‖ = ‖fderiv 𝕜 f x‖

    TauCeti.Analysis.Calculus.Gradient · lines 72–73

    Why only part: The declaration fixes the size of the gradient as the size of the differential at the point, and leaves open which direction realises that size.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 0db818cdca86 · pair key 6f03abdd946b.
  4. Definition · main statement

    Partial

    Partial: TauCeti states a special case (a definition).

    Further, a point where the gradient is the zero vector is known as a stationary point.

    Show in article · Read it in context on Wikipedia

    TauCeti.IsNondegenerateCriticalPoint special case

    structure TauCeti.IsNondegenerateCriticalPoint {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] : (E → ℝ) → E → Prop

    TauCeti.Analysis.Calculus.Morse.Basic · lines 124–130

    Why a special case: The declaration singles out the points with vanishing differential that also carry two regularity conditions, so it names the passage's notion within the Morse setting.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key b87d54c4f9df · pair key a4da0907eae8.
  5. Definition · main statement

    Partial

    Partial: Mathlib states part of this (proved).

    In coordinate-free terms, the gradient of a function f(r) may be defined by: df = ∇f ⋅ dr where df is the total infinitesimal change in f for an infinitesimal displacement dr, and is seen to be maximal when dr is in the direction of the gradient ∇f.

    Show in article · Read it in context on Wikipedia

    inner_gradient_left states part

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why only part: Mathlib states this identity for any complete inner product space over a real or complex scalar field, writing the total change of f as the Fréchet derivative applied to the displacement.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key aac9c94426c6 · pair key a7aba8d56b7c.
  6. Notation · supporting

    Partial

    Partial: Mathlib states a special case (a definition).

    The nabla symbol ∇, written as an upside-down triangle and pronounced "del", denotes the vector differential operator.

    Show in article · Read it in context on Wikipedia

    Gradient.«term∇» special case

    def Gradient.«term∇» : Lean.ParserDescr

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 86–87

    Why a special case: Mathlib introduces the symbol inside its Gradient namespace so that writing it abbreviates the gradient operator.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 788a2ab16607 · pair key 54d42e97ee9e.
  7. Definition · main statement

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    When a coordinate system is used in which the basis vectors are not functions of position, the gradient is given by the vector[a] whose components are the partial derivatives of f at p.[2]

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib builds the gradient from the inner product rather than from coordinates, and in a basis that is fixed and orthonormal its entries come out as the partial derivatives of f.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key dfe0284eef63 · pair key 3b2826bccd1c.
  8. Definition · main statement

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    That is, for f: ℝn → ℝ, its gradient ∇f: ℝn → ℝn is defined at the point p = (x1, …, xn) in n-dimensional space as the vector[b] ∇f(p) = [∂f∂x1(p)⋮∂f∂xn(p)].

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib defines the gradient for scalar functions on any complete inner product space, and over the standard inner product of n-dimensional real space its entries are the partial derivatives of f with respect to each coordinate.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 0d176537a7eb · pair key 3174acf3d74c.
  9. Theorem · main statement

    Not formalized
    There can be functions for which partial derivatives exist in every direction but fail to be differentiable.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 6d36f9c76983).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  10. Example · supporting

    Not formalized
    For example, the function f(x, y) = x2yx2 + y2 unless at origin where f(0, 0) = 0, is not differentiable at the origin as it does not have a well defined tangent plane despite having well defined partial derivatives in every direction at the origin.[3]

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt d30ab259b6c4).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  11. Example · supporting

    Not formalized
    In this particular example, under rotation of x-y coordinate system, the above formula for gradient fails to transform like a vector (gradient becomes dependent on choice of basis for coordinate system) and also fails to point towards the 'steepest ascent' in some orientations.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 2 candidates examined (receipt e8054f5f699d).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  12. Theorem · main statement

    Not formalized
    For differentiable functions where the formula for gradient holds, it can be shown to always transform as a vector under transformation of the basis so as to always point towards the fastest increase.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 32435375f742).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  13. Theorem · main statement

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The gradient is dual to the total derivative df: the value of the gradient at a point is a tangent vector – a vector at each point; while the value of the derivative at a point is a cotangent vector – a linear functional on vectors.[c]

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient_def more general

    theorem TauCeti.Manifold.mgradient_def {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] (f : M → ℝ) (x : M) : TauCeti.Manifold.mgradient I f x = (Riemannian.Tensor.rieszDual x) (d% f x)

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 63–65

    Why more general: The declaration is stated for an arbitrary Riemannian manifold; its docstring calls the gradient the fibrewise Riesz dual of the manifold derivative, which is what the passage asserts.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 7a6bf77cc637 · pair key 3be310e60877.

    hasGradientAt_iff_hasFDerivAt more general

    theorem hasGradientAt_iff_hasFDerivAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x : F} : HasGradientAt f f' x ↔ HasFDerivAt f ((InnerProductSpace.toDual 𝕜 F) f') x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 103–105

    Why more general: Written over any complete inner product space, with the Riesz map toDual carrying the vector to the functional and back.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key a72971d11174 · pair key 1629b7a70624.

    TauCeti.Manifold.mgradient more general

    def TauCeti.Manifold.mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] (I : ModelWithCorners ℝ E H) {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] : (M → ℝ) → (x : M) → TangentSpace I x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 59–60

    Community Agent The pipeline said states part; changed by Meta FAIR 2 (agent) — owned by Jack McCarthy on 10 Nov 2026.

    Why more general: TauCeti.Manifold.mgradient defines the gradient at x as the Riesz dual, through the Riemannian metric, of the differential of f at x. Its value is a tangent vector, and the differential it dualises is a linear functional on tangent vectors. That is every clause of the passage, stated for any Riemannian manifold, of which Euclidean space is a case.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key afa0c8b75bf1 · pair key cddacbff8727.
  14. Theorem · main statement

    Formalized

    Formalized: Mathlib states a more general result (proved).

    They are related in that the dot product of the gradient of f at a point p with another tangent vector v equals the directional derivative of f at p of the function along v; that is, ∇f(p) ⋅ v = ∂f∂v(p) = dfp(v).

    Show in article · Read it in context on Wikipedia

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: Both sides fall back on zero where f has no derivative, so the case outside the passage's differentiable setting is a default value with no content.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key b5a17b24003f · pair key 071424947a44.

    TauCeti.Manifold.inner_mgradient more general

    theorem TauCeti.Manifold.inner_mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] {f : M → ℝ} {x : M} (v : TangentSpace I x) : Inner.inner ℝ (TauCeti.Manifold.mgradient I f x) v = (d% f x) v

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 71–73

    Why more general: Stated on an arbitrary Riemannian manifold, where the metric plays the role the dot product plays in the passage.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 7c43ce778868 · pair key ca656590fc17.
  15. Example · supporting · Motivation

    Not formalized
    Consider a room where the temperature is given by a scalar field, T, so at each point (x, y, z) the temperature is T(x, y, z), independent of time. At each point in the room, the gradient of T at that point will show the direction in which the temperature rises most quickly, moving away from (x, y, z).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 734d045e0041).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  16. Example · supporting · Motivation

    Not formalized
    The magnitude of the gradient will determine how fast the temperature rises in that direction.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 05629e412476).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  17. Example · supporting · Motivation

    Not formalized
    Consider a surface whose height above sea level at point (x, y) is H(x, y). The gradient of H at a point is a plane vector pointing in the direction of the steepest slope or grade at that point.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 4c7181e2b207).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  18. Example · supporting · Motivation

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The steepness of the slope at that point is given by the magnitude of the gradient vector.

    Show in article · Read it in context on Wikipedia

    TauCeti.norm_gradient_eq_norm_fderiv more general

    theorem TauCeti.norm_gradient_eq_norm_fderiv {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] (f : F → 𝕜) (x : F) : ‖gradient f x‖ = ‖fderiv 𝕜 f x‖

    TauCeti.Analysis.Calculus.Gradient · lines 72–73

    Why more general: Lean measures the steepness at the point as the operator norm of the differential, and the lemma equates that number with the length of the gradient vector.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 6b2ecda8e3ab · pair key 646c97200396.
  19. Example · supporting · Motivation

    Not formalized
    Suppose that the steepest slope on a hill is 40%. A road going directly uphill has slope 40%, but a road going around the hill at an angle will have a shallower slope.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt fcc062a7fda1).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  20. Example · supporting · Motivation

    Not formalized
    For example, if the road is at a 60° angle from the uphill direction (when both directions are projected onto the horizontal plane), then the slope along the road will be the dot product between the gradient vector and a unit vector along the road, as the dot product measures how much the unit vector along the road aligns with the steepest slope,[d] which is 40% times the cosine of 60°, or 20%.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt fe3cb75f3d12).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  21. Theorem · main statement · Motivation

    Formalized

    Formalized: Mathlib states a more general result (proved).

    More generally, if the hill height function H is differentiable, then the gradient of H dotted with a unit vector gives the slope of the hill in the direction of the vector, the directional derivative of H along the unit vector.

    Show in article · Read it in context on Wikipedia

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: Lean writes the derivative of H along a heading as the Fréchet derivative applied to it, and the lemma holds for every vector, unit length or not.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5559e16feb5b · pair key e3b91fa83555.

    HasGradientAt.fderiv_apply more general

    theorem HasGradientAt.fderiv_apply {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x y : F} : HasGradientAt f f' x → (fderiv 𝕜 f x) y = Inner.inner 𝕜 f' y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 282–284

    Why more general: The hypothesis names the gradient as a witness, which for a differentiable height function is the gradient the passage writes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 553a20d3f147 · pair key 0cee0aa3633d.
  22. Notation · supporting · Notation

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The gradient of a function f at point a is usually written as ∇f(a). It may also be denoted by any of the following:

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Its docstring gives ∇ as the symbol for this operator inside the Gradient namespace, and the declaration itself carries the name gradient.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 3bfc9121c34f · pair key fdec7f378002.
  23. Notation · supporting · Notation

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    It may also be denoted by any of the following:
    ∇⃗f(a): to emphasize the vector nature of the result.

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Lean writes the symbol without an arrow, and the value gradient f a lives in the space F, so it is a vector, which is what the arrow is there to say.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 295ee9217294 · pair key 154fd802b4aa.
  24. Notation · supporting · Notation

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    grad f

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Lean names the operator gradient and offers ∇ as a scoped symbol for it, so grad f is gradient f.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 9c08e67a52e4 · pair key cb1f5f06ddfc.
  25. Notation · supporting · Notation

    Formalized

    Formalized: ATLAS (atlas-lean) states this up to a convention difference (a definition).

    grad f
    ∂if and fi: Written with Einstein notation, where repeated indices (i) are summed over.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.partialDeriv close match

    def MetaMathlibExt.partialDeriv {n : ℕ} : ((Fin n → ℝ) → ℝ) → Fin n → (Fin n → ℝ) → ℝ

    MathlibExt.Analysis.Calculus.HigherPartials · line 37

    Why close, not exact: Lean names the coordinate derivative along the i-th axis by an ordinary application of a function to f and to i, where the passage writes a subscripted index and leaves repetition of that index to the summation convention of the reading.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key c8bad1c12843 · pair key e24277f2dfe0.
  26. Notation · supporting · Definition

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The gradient (or gradient vector field) of a scalar function f(x1, x2, x3, …, xn) is denoted ∇f or ∇→f where ∇ (nabla) denotes the vector differential operator, del.

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib defines this gradient vector field for a scalar function on any complete inner product space over the reals or complexes and writes it with the nabla sign inside its gradient namespace, so the passage's function of n real variables is one reading of it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 712f3d2f30b4 · pair key 7a5f7ebfc491.
  27. Notation · supporting · Definition

    Not formalized

    With community edits: Not formalized. As published it is Formalized.

    The notation grad f is also commonly used to represent the gradient.

    Show in article · Read it in context on Wikipedia

  28. Definition · main statement · Definition

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The gradient of f is defined as the unique vector field whose dot product with any vector v at each point x is the directional derivative of f along v. That is, (∇f(x)) ⋅ v̂ = Dvf(x) where the right-hand side is the directional derivative and there are many ways to represent it.

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib builds this vector as the Riesz representative of the derivative of f, so it is by construction the one vector whose pairing with each direction returns the derivative of f in that direction, and it does this for a scalar function on any complete inner product space over the reals or complexes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key df9c4eccde82 · pair key 5a138958040b.

    inner_gradient_left states part

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why only part: It supplies the pairing identity of the passage's definition, with the directional derivative written as the derivative map applied to the direction.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key e27d8c929532 · pair key c9d24759c8dd.

    HasGradientAt.unique states part

    theorem HasGradientAt.unique {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x gradf gradg : F} : HasGradientAt f gradf x → HasGradientAt f gradg x → gradf = gradg

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 140–143

    Why only part: It supplies the uniqueness half of the passage's description of the gradient as the one vector field with the stated pairing property.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 93d21674b050 · pair key 2b819269ea89.
  29. Theorem · main statement · Definition

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Formally, the derivative is dual to the gradient; see relationship with derivative.

    Show in article · Read it in context on Wikipedia

    toDual_gradient more general

    theorem toDual_gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : (InnerProductSpace.toDual 𝕜 F) (gradient f x) = fderiv 𝕜 f x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 128–129

    Why more general: Mathlib states the duality as an equation: the Riesz duality map sends the gradient vector at a point to the derivative functional at that point, over any complete inner product space over the reals or complexes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key cde16c23b3e8 · pair key 2ff8bca60e06.

    hasGradientAt_iff_hasFDerivAt more general

    theorem hasGradientAt_iff_hasFDerivAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x : F} : HasGradientAt f f' x ↔ HasFDerivAt f ((InnerProductSpace.toDual 𝕜 F) f') x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 103–105

    Why more general: Mathlib states the duality as an equivalence between having a gradient vector and having the corresponding dual functional as derivative, over any complete inner product space over the reals or complexes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 43185259f063 · pair key bac897133e4c.

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: Mathlib writes the duality pointwise: applying the derivative to a vector is the same as pairing that vector with the gradient, for a scalar function on any complete inner product space over the reals or complexes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 9ff1205d2dcc · pair key de72382facfc.
  30. Notation · supporting · Definition

    Partial

    Partial: ATLAS (atlas-lean) states a special case (a definition).

    When a function also depends on a parameter such as time, the gradient often refers simply to the vector of its spatial derivatives only (see Spatial gradient).

    Show in article · Read it in context on Wikipedia

    MathlibExt.Analysis.FiniteTimeSingularityWanted.FiniteTimeSingularity.spaceGrad special case

    def MathlibExt.Analysis.FiniteTimeSingularityWanted.FiniteTimeSingularity.spaceGrad : (EuclideanSpace ℝ (Fin 3) → ℝ → ℝ) → EuclideanSpace ℝ (Fin 3) → ℝ → EuclideanSpace ℝ (Fin 3)

    WantedExt.Analysis.FiniteTimeSingularityWanted · line 117

    Why a special case: The declaration fixes three space dimensions and takes the parameter to be a real time, where the passage speaks of a function of any number of space variables and any parameter.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 8893ee1a1267 · pair key 6c64f1580811.
  31. Theorem · main statement · Definition

    Not formalized
    The magnitude and direction of the gradient vector are independent of the particular coordinate representation.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 9b5db47b720a).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  32. Theorem · main statement · Definition › Cartesian coordinates

    Not formalized
    In the three-dimensional Cartesian coordinate system with a Euclidean metric, the gradient, if it exists, is given by ∇f = ∂f∂xi + ∂f∂yj + ∂f∂zk, where i, j, k are the standard unit vectors in the directions of the x, y and z coordinates, respectively.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 218264cc39eb).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  33. Example · supporting · Definition › Cartesian coordinates

    Not formalized
    For example, the gradient of the function f(x, y, z) = 2x + 3y2 − sin(z) is ∇f(x, y, z) = 2i + 6yj − cos(z)k. or ∇f(x, y, z) = [26y−cos z].

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 5cd5f33d37eb).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  34. Theorem · main statement · Definition › Cylindrical and spherical coordinates

    Not formalized
    In cylindrical coordinates, the gradient is given by:[6] ∇f(ρ, φ, z) = ∂f∂ρeρ + 1ρ∂f∂φeφ + ∂f∂zez, where ρ is the axial distance, φ is the azimuthal or azimuth angle, z is the axial coordinate, and eρ, eφ and ez are unit vectors pointing along the coordinate directions.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 1956861c02ba).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  35. Theorem · main statement · Definition › Cylindrical and spherical coordinates

    Not formalized
    In spherical coordinates with a Euclidean metric, the gradient is given by:[6] ∇f(r, θ, φ) = ∂f∂rer + 1r∂f∂θeθ + 1rsin θ∂f∂φeφ, where r is the radial distance, φ is the azimuthal angle and θ is the polar angle, and er, eθ and eφ are again local unit vectors pointing in the coordinate directions (that is, the normalized covariant basis).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 1 candidates examined (receipt 8b5cbfafef4e).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  36. Theorem · main statement · Definition › General coordinates

    Not formalized
    Using Einstein notation, the gradient can then be written as:

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 2 candidates examined (receipt 57c4a68bb5f2).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  37. Theorem · main statement · Definition › General coordinates

    Formalized

    Formalized by 2 parts together (more general):

    1. Riemannian.Tensor.rieszDual_eq_sum_chartLocalFrame TauCeti (proved)
      • in general coordinates the gradient of f equals the partial derivative of f with respect to x^i times the inverse metric component g^{ij} times the unnormalized basis vector e_j, summed over i and j
    2. Module.Basis.sum_dual_apply_smul_coord Mathlib (proved)
      • the differential dual to that gradient equals the partial derivative of f with respect to x^i times the dual basis covector e^i, summed over i
    ∇f = ∂f∂xigijej (Note that its dual is df = ∂f∂xiei),

    Show in article · Read it in context on Wikipedia

    Riemannian.Tensor.rieszDual_eq_sum_chartLocalFrame states part

    theorem Riemannian.Tensor.rieszDual_eq_sum_chartLocalFrame {E : Type*} […] {H : Type*} {I : ModelWithCorners ℝ E H} {M : Type*} (a : M) {x : M} : x ∈ (chartAt H a).source → ∀ (φ : TangentSpace I x →L[ℝ] ℝ), (Riemannian.Tensor.rieszDual x) φ = ∑ i, (Riemannian.Tensor.chartInvGramMatrix a x).mulVec (fun j ↦ φ (Riemannian.Tensor.chartLocalFrame a j x)) i • Riemannian.Tensor.chartLocalFrame a i x
    Full type, as Lean prints it
    ∀ {E : Type u_1} [inst : NormedAddCommGroup E] [inst_1 : NormedSpace ℝ E] [inst_2 : FiniteDimensional ℝ E] {H : Type u_2} [inst_3 : TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type u_3} [inst_4 : TopologicalSpace M] [inst_5 : ChartedSpace H M] [inst_6 : IsManifold I 1 M] [inst_7 : Bundle.RiemannianBundle fun x ↦ TangentSpace I x] (a : M) {x : M}, x ∈ (chartAt H a).source → ∀ (φ : TangentSpace I x →L[ℝ] ℝ), (Riemannian.Tensor.rieszDual x) φ = ∑ i, (Riemannian.Tensor.chartInvGramMatrix a x).mulVec (fun j ↦ φ (Riemannian.Tensor.chartLocalFrame a j x)) i • Riemannian.Tensor.chartLocalFrame a i x

    TauCeti.Geometry.Manifold.VectorBundle.Riemannian.Riesz · lines 99–136

    Why only part: The declaration is written for an arbitrary cotangent covector, of which the differential of f is one, and the passage's gradient is by the article's own definition the metric dual of that differential.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 322aef12055f · pair key 8ce34e4a5493.

    Module.Basis.sum_dual_apply_smul_coord states part

    theorem Module.Basis.sum_dual_apply_smul_coord {R : Type uR} {M : Type uM} {ι : Type uι} [CommSemiring R] [AddCommMonoid M] [_root_.Module R M] (b : Module.Basis ι R M) [Fintype ι] (f : Module.Dual R M) : ∑ x, f (b x) • b.coord x = f

    Mathlib.LinearAlgebra.Dual.Basis · lines 112–116

    Why only part: The declaration is a basis expansion valid over any commutative semiring and any finite basis; the passage reads it in one tangent space with the coordinate frame and its dual covectors.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 08d246cb6e0e · pair key 42f26e6e11d1.
  38. Notation · supporting · Definition › General coordinates

    Partial

    Partial: TauCeti states part of this (a definition).

    where ei = dxi and ei = ∂x/∂xi refer to the unnormalized local covariant and contravariant bases respectively, gij is the inverse metric tensor, and the Einstein summation convention implies summation over i and j.

    Show in article · Read it in context on Wikipedia

    Riemannian.Tensor.chartLocalFrame states part

    def Riemannian.Tensor.chartLocalFrame {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] : M → Fin (Module.finrank ℝ E) → (x : M) → TangentSpace I x

    TauCeti.Geometry.Manifold.VectorBundle.Riemannian.ChartGram · lines 72–74

    Why only part: The declaration names the unnormalized coordinate frame the passage writes as the derivative of the position with respect to x^i; it covers neither the dual covectors nor the inverse metric nor the summation convention.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 5697bdcb73ef · pair key b5fc85779f92.

    Riemannian.Tensor.chartInvGramMatrix states part

    def Riemannian.Tensor.chartInvGramMatrix {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] : M → M → Matrix (Fin (Module.finrank ℝ E)) (Fin (Module.finrank ℝ E)) ℝ

    TauCeti.Geometry.Manifold.VectorBundle.Riemannian.ChartGram · lines 230–232

    Why only part: Its docstring calls it the inverse coordinate-metric matrix, which is what the passage means by the inverse metric tensor; it says nothing about the two bases or the summation convention.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 59c1a0a0ef46 · pair key 62f51c304e8f.
  39. Definition · main statement · Definition › General coordinates

    Not formalized
    If the coordinates are orthogonal we can easily express the gradient (and the differential) in terms of the normalized bases, which we refer to as êi and êi, using the scale factors (also known as Lamé coefficients) hi = ‖ei‖ = √gii = 1 /‖ei‖:

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 8d1ff1a9fb9a).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  40. Theorem · main statement · Definition › General coordinates

    Not formalized
    ∇f = ∂f∂xigijêj√gjj = ∑ni = 1 ∂f∂xi1hiêi (and df = ∑ni = 1 ∂f∂xi1hiêi), where we cannot use Einstein notation, since it is impossible to avoid the repetition of more than two indices.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 3 candidates examined (receipt e50fa7c1910a).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  41. Theorem · supporting · Definition › General coordinates

    Not formalized
    Despite the use of upper and lower indices, êi, êi, and hi are neither contravariant nor covariant.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt ff0d859fb0e3).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  42. Theorem · supporting · Definition › General coordinates

    Not formalized
    The latter expression evaluates to the expressions given above for cylindrical and spherical coordinates.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 07463ac43f95).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  43. Theorem · main statement · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The gradient is closely related to the total derivative (total differential) df: they are transpose (dual) to each other.

    Show in article · Read it in context on Wikipedia

    toDual_gradient more general

    theorem toDual_gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : (InnerProductSpace.toDual 𝕜 F) (gradient f x) = fderiv 𝕜 f x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 128–129

    Why more general: The declaration holds in any complete inner product space over a real or complex field, and over real n-space the canonical duality map is exactly the transposition the passage describes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 47d7fc5f5ced · pair key 2c387a727fd5.

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: The declaration is stated over any complete inner product space above a real or complex field; the passage reads it over real n-space, where pairing a vector against the dot product is transposition.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 67a4ef3b5c97 · pair key 2d3fd0501155.
  44. Theorem · main statement · Relationship with derivative › Relationship with total derivative

    Partial

    Partial: Mathlib states part of this (proved).

    Using the convention that vectors in ℝn are represented by column vectors, and that covectors (linear maps ℝn → ℝ) are represented by row vectors,[a] the gradient ∇f and the derivative df are expressed as a column and row vector, respectively, with the same components, but transpose of each other: ∇f(p) = [∂f∂x1(p)⋮∂f∂xn(p)]; dfp = [∂f∂x1(p)⋯∂f∂xn(p)].

    Show in article · Read it in context on Wikipedia

    inner_gradient_left states part

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why only part: The declaration holds over any complete inner product space above a real or complex field, and over real n-space with the dot product the pairing against a coordinate vector is the entry of the column vector the passage describes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 764cf65e54af · pair key fff6d5b869e0.

    MetaMathlibExt.partialDeriv_eq_fderiv_apply states part

    theorem MetaMathlibExt.partialDeriv_eq_fderiv_apply {n : ℕ} (f : (Fin n → ℝ) → ℝ) (i : Fin n) (x : Fin n → ℝ) : DifferentiableAt ℝ f x → MetaMathlibExt.partialDeriv f i x = (fderiv ℝ f x) (Pi.single i 1)

    MathlibExt.Analysis.Calculus.HigherPartials · line 66

    Why only part: It names the entries of the derivative covector but never mentions the gradient, so it leaves the transposition half of this conjunct unstated.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 185cbbaa0f3a · pair key 42260a6726d1.

    toDual_gradient states part

    theorem toDual_gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : (InnerProductSpace.toDual 𝕜 F) (gradient f x) = fderiv 𝕜 f x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 128–129

    Why only part: It states that the derivative is the dual of the gradient but names no entries, so it does not by itself give the row vector of partial derivatives.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ab8fffca9b4a · pair key 0b351bbbeab9.
  45. Theorem · supporting · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized by 2 parts together (more general):

    1. fderiv Mathlib (definition)
      • at each point the total derivative of f is a cotangent vector, that is a linear form giving the change in the scalar output for an infinitesimal change of the vector input
    2. gradient Mathlib (definition)
      • at each point the gradient of f is a tangent vector, which represents an infinitesimal change of the vector input
    While these both have the same components, they differ in what kind of mathematical object they represent: at each point, the derivative is a cotangent vector, a linear form (or covector) which expresses how much the (scalar) output changes for a given infinitesimal change in (vector) input, while at each point, the gradient is a tangent vector, which represents an infinitesimal change in (vector) input.

    Show in article · Read it in context on Wikipedia

    fderiv states part

    def fderiv (𝕜 : Type*) [NontriviallyNormedField 𝕜] {E : Type*} [AddCommGroup E] [_root_.Module 𝕜 E] [TopologicalSpace E] {F : Type*} [AddCommGroup F] [_root_.Module 𝕜 F] [TopologicalSpace F] : (E → F) → E → E →L[𝕜] F

    Mathlib.Analysis.Calculus.FDeriv.Defs · lines 160–161

    Why only part: Lean's value is a continuous linear map, and on a finite-dimensional input space continuity is automatic for linear maps; the first-order approximation recorded in its docstring is the infinitesimal change of output the sentence describes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6be790fcf7bf · pair key fe7d87c7b8c4.

    gradient states part

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why only part: Its docstring records the first-order expansion of the function through the inner product with this vector, which is the infinitesimal change of input the sentence means.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 09525c1bf15b · pair key 472d92d74506.
  46. Theorem · supporting · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized by 2 parts together (more general):

    1. TauCeti.Manifold.mgradient TauCeti (definition)
      • the gradient of f at p lies in the tangent space at p of n-dimensional real space
    2. mfderiv Mathlib (definition)
      • the total derivative of f at p is a map from the tangent space at p to the real numbers
    In symbols, the gradient is an element of the tangent space at a point, ∇f(p) ∈ Tpℝn, while the derivative is a map from the tangent space to the real numbers, dfp: Tpℝn → ℝ.

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient states part

    def TauCeti.Manifold.mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] (I : ModelWithCorners ℝ E H) {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] : (M → ℝ) → (x : M) → TangentSpace I x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 59–60

    Why only part: The declaration is written for a Riemannian manifold, and the article's space with its usual inner product is one of those; its docstring describes the value as the tangent vector dual to the differential.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 88e8c8ce3ddf · pair key cbd8fce8610c.

    mfderiv states part

    def mfderiv {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {H : Type*} [TopologicalSpace H] (I : ModelWithCorners 𝕜 E H) {M : Type*} [TopologicalSpace M] [ChartedSpace H M] {E' : Type*} [NormedAddCommGroup E'] [NormedSpace 𝕜 E'] {H' : Type*} [TopologicalSpace H'] (I' : ModelWithCorners 𝕜 E' H') {M' : Type*} [TopologicalSpace M'] [ChartedSpace H' M'] (f : M → M') (x : M) : TangentSpace I x →L[𝕜] TangentSpace I' (f x)

    Mathlib.Geometry.Manifold.MFDeriv.Defs · lines 336–341

    Why only part: Its docstring describes the value as the derivative seen as a continuous linear map between tangent spaces, and the tangent space of the reals at any point is the reals.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 1b0b982ea68c · pair key 488d91733f23.
  47. Theorem · supporting · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The tangent spaces at each point of ℝn can be "naturally" identified[e] with the vector space ℝn itself, and similarly the cotangent space at each point can be naturally identified with the dual vector space (ℝn)∗ of covectors; thus the value of the gradient at a point can be thought of a vector in the original ℝn, not just as a tangent…

    Show in article · Read it in context on Wikipedia

    NormedSpace.fromTangentSpace more general

    def NormedSpace.fromTangentSpace {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] (v : E) : TangentSpace (modelWithCornersSelf 𝕜 E) v ≃L[𝕜] E

    Mathlib.Geometry.Manifold.IsManifold.Basic · lines 1063–1064

    Why more general: Its docstring calls this the canonical identification, and the value is a linear homeomorphism, which carries the vector-space identification the sentence claims.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5e587da2ea31 · pair key 4a746b4f9a4d.
  48. Theorem · supporting · Relationship with derivative › Relationship with total derivative

    Not formalized
    The tangent spaces at each point of ℝn can be "naturally" identified[e] with the vector space ℝn itself, and similarly the cotangent space at each point can be naturally identified with the dual vector space (ℝn)∗ of covectors; thus the value of the gradient at a point can be thought of a vector in the original ℝn, not just as a tangent vector.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 3 candidates examined (receipt f96c1c658a85).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  49. Theorem · supporting · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized: TauCeti states a more general result (proved).

    …each point of ℝn can be "naturally" identified[e] with the vector space ℝn itself, and similarly the cotangent space at each point can be naturally identified with the dual vector space (ℝn)∗ of covectors; thus the value of the gradient at a point can be thought of a vector in the original ℝn, not just as a tangent vector.

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient_eq_gradient more general

    theorem TauCeti.Manifold.mgradient_eq_gradient {F : Type*} [NormedAddCommGroup F] [InnerProductSpace ℝ F] [FiniteDimensional ℝ F] (f : F → ℝ) (x : F) : have x_1 := { g := (riemannianMetricVectorSpace F).toRiemannianMetric }; TauCeti.Manifold.mgradient (modelWithCornersSelf ℝ F) f x = gradient f x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.NormedSpace · lines 44–61

    Why more general: Its two sides are the gradient taken as a tangent vector and the gradient taken as a vector of the underlying space, so the equation is the identification the sentence invokes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key a4350eaa3b64 · pair key d9b9efa6b967.
  50. Theorem · main statement · Relationship with derivative › Relationship with total derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Computationally, given a tangent vector, the vector can be multiplied by the derivative (as matrices), which is equal to taking the dot product with the gradient: (dfp)(v) = [∂f∂x1(p)⋯∂f∂xn(p)][v1⋮vn] = ∑ni = 1 ∂f∂xi(p)vi = [∂f∂x1(p)⋮∂f∂xn(p)] ⋅ [v1⋮vn] = ∇f(p) ⋅ v

    Show in article · Read it in context on Wikipedia

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: Over the reals the inner product is symmetric, so the side carrying the gradient does not matter; the sum over partial derivatives in the article is the middle step of its chain rather than a further claim.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 191cf843f079 · pair key c496ab1b5b1b.

    HasGradientAt.fderiv_apply more general

    theorem HasGradientAt.fderiv_apply {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x y : F} : HasGradientAt f f' x → (fderiv 𝕜 f x) y = Inner.inner 𝕜 f' y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 282–284

    Why more general: It states the same identity for an arbitrary witness of the gradient relation instead of for the chosen value of the gradient.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 168df87b8906 · pair key 7b45955b407a.
  51. Definition · main statement · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The best linear approximation to a differentiable function f : ℝn → ℝ at a point x in ℝn is a linear map from ℝn to ℝ which is often denoted by dfx or Df(x) and called the differential or total derivative of f at x.

    Show in article · Read it in context on Wikipedia

    fderiv more general

    def fderiv (𝕜 : Type*) [NontriviallyNormedField 𝕜] {E : Type*} [AddCommGroup E] [_root_.Module 𝕜 E] [TopologicalSpace E] {F : Type*} [AddCommGroup F] [_root_.Module 𝕜 F] [TopologicalSpace F] : (E → F) → E → E →L[𝕜] F

    Mathlib.Analysis.Calculus.FDeriv.Defs · lines 160–161

    Why more general: Its docstring says the value is a derivative of the function at the point whenever one exists, which is the approximating linear map the article names and denotes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d16d3fb700cb · pair key 9e2a898e93a7.
  52. Definition · supporting · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Partial

    Partial: Mathlib states a more general result, but as a proved theorem, not a definition; a definition passage is formalized only by a Lean definition.

    The function df, which maps x to dfx, is called the total differential or exterior derivative of f and is an example of a differential 1-form.

    Show in article · Read it in context on Wikipedia

    extDeriv_constOfIsEmpty more general

    theorem extDeriv_constOfIsEmpty {𝕜 : Type*} {E : Type*} {F : Type*} [NontriviallyNormedField 𝕜] [NormedAddCommGroup E] [NormedSpace 𝕜 E] [NormedAddCommGroup F] [NormedSpace 𝕜 F] (f : E → F) (x : E) : extDeriv (fun x ↦ ContinuousAlternatingMap.constOfIsEmpty 𝕜 E (Fin 0) (f x)) x = (ContinuousAlternatingMap.ofSubsingleton 𝕜 E F 0) (fderiv 𝕜 f x)

    Mathlib.Analysis.Calculus.DifferentialForm.Basic · lines 144–147

    Why more general: Lean presents the function as a form of degree zero and the differential as a form of degree one through the canonical maps for an empty and a one-element index, and the value lives in the type the library uses for differential forms of degree one.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 238d89e206c6 · pair key 9c7fd8364443.
  53. Theorem · supporting · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Partial

    Partial: Mathlib states part of this (proved).

    Much as the derivative of a function of a single variable represents the slope of the tangent to the graph of the function,[7] the directional derivative of a function in several variables represents the slope of the tangent hyperplane in the direction of the vector.

    Show in article · Read it in context on Wikipedia

    hasLineDerivAt_iff_tendsto_slope_zero states part

    theorem hasLineDerivAt_iff_tendsto_slope_zero {𝕜 : Type*} [NontriviallyNormedField 𝕜] {F : Type*} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {E : Type*} [AddCommGroup E] [_root_.Module 𝕜 E] {f : E → F} {f' : F} {x v : E} : HasLineDerivAt 𝕜 f f' x v ↔ Filter.Tendsto (fun t ↦ t⁻¹ • (f (x + t • v) - f x)) (𝓝[≠] 0) (𝓝 f')

    Mathlib.Analysis.Calculus.LineDeriv.Basic · lines 213–217

    Why only part: the library names the limit of difference quotients along v; the tangent hyperplane whose slope the passage speaks of is not an object of the index

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ce61c2626a99 · pair key baae3fa74ea4.

    hasDerivAt_iff_tendsto_slope states part

    theorem hasDerivAt_iff_tendsto_slope {𝕜 : Type u} [NontriviallyNormedField 𝕜] {F : Type v} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {f : 𝕜 → F} {f' : F} {x : 𝕜} : HasDerivAt f f' x ↔ Filter.Tendsto (slope f x) (𝓝[≠] x) (𝓝 f')

    Mathlib.Analysis.Calculus.Deriv.Slope · lines 73–74

    Why only part: this covers the one-variable comparison the passage draws and says nothing about several variables

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 2d0abb3cab26 · pair key a3c23a2cbc87.
  54. Theorem · main statement · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The gradient is related to the differential by the formula (∇f)x ⋅ v = dfx(v) for any v ∈ ℝn, where ⋅ is the dot product: taking the dot product of a vector with the gradient is the same as taking the directional derivative along the vector.

    Show in article · Read it in context on Wikipedia

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: it is stated for every complete inner product space over the reals or the complex numbers, and the passage's space is one of them

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ad96d8de7dca · pair key 7c33ca8f408f.

    HasGradientAt.fderiv_apply more general

    theorem HasGradientAt.fderiv_apply {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x y : F} : HasGradientAt f f' x → (fderiv 𝕜 f x) y = Inner.inner 𝕜 f' y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 282–284

    Why more general: the equation is written with the differential on the left, which is the passage's reading from the other side

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 734b32096264 · pair key 757e448e489c.
  55. Theorem · supporting · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Partial

    Partial: ATLAS (atlas-lean) states part of this (proved).

    If ℝn is viewed as the space of (dimension n) column vectors (of real numbers), then one can regard df as the row vector with components (∂f∂x1, …, ∂f∂xn), so that dfx(v) is given by matrix multiplication.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.partialDeriv_eq_fderiv_apply states part

    theorem MetaMathlibExt.partialDeriv_eq_fderiv_apply {n : ℕ} (f : (Fin n → ℝ) → ℝ) (i : Fin n) (x : Fin n → ℝ) : DifferentiableAt ℝ f x → MetaMathlibExt.partialDeriv f i x = (fderiv ℝ f x) (Pi.single i 1)

    MathlibExt.Analysis.Calculus.HigherPartials · line 66

    Why only part: it pins the entries of the differential and leaves the matrix-product form of the differential applied to a vector unstated

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 09e783228a7a · pair key 8e20a285f949.
  56. Theorem · supporting · Relationship with derivative › Relationship with total derivative › Differential or (exterior) derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Assuming the standard Euclidean metric on ℝn, the gradient is then the corresponding column vector, that is, (∇f)i = dfTi.

    Show in article · Read it in context on Wikipedia

    toDual_gradient more general

    theorem toDual_gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : (InnerProductSpace.toDual 𝕜 F) (gradient f x) = fderiv 𝕜 f x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 128–129

    Why more general: the correspondence is written through the metric's duality map rather than as a transpose of coordinates; with the standard dot product on the passage's space the two descriptions pick out the same vector

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 2e233935cfc6 · pair key baf607b9317f.

    inner_gradient_left more general

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Why more general: this is the same correspondence written on vectors instead of on the dual side

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key baaf7ac23ff2 · pair key f7322b809c62.
  57. Theorem · main statement · Relationship with derivative › Relationship with total derivative › Linear approximation to a function

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The gradient of a function f from the Euclidean space ℝn to ℝ at any particular point x0 in ℝn characterizes the best linear approximation to f at x0.

    Show in article · Read it in context on Wikipedia

    hasGradientAt_iff_isLittleO more general

    theorem hasGradientAt_iff_isLittleO {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x : F} : HasGradientAt f f' x ↔ (fun x' ↦ f x' - f x - Inner.inner 𝕜 f' (x' - x)) =o[𝓝 x] fun x' ↦ x' - x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 248–250

    Why more general: the statement is two-way, which is what the passage's word characterizes asks for, and it is written for any complete real or complex inner product space

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key ae49348b67e3 · pair key 8f07d8bffdbd.

    HasGradientAt.unique states part

    theorem HasGradientAt.unique {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x gradf gradg : F} : HasGradientAt f gradf x → HasGradientAt f gradg x → gradf = gradg

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 140–143

    Why only part: it supplies the uniqueness half of the passage's claim that the gradient pins the approximation down

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key f4b39957c34c · pair key ad031cd56989.
  58. Theorem · main statement · Relationship with derivative › Relationship with total derivative › Linear approximation to a function

    Formalized

    Formalized: Mathlib states a more general result (proved).

    The approximation is as follows: f(x) ≈ f(x0) + (∇f)x0 ⋅ (x − x0) for x close to x0, where (∇f)x0 is the gradient of f computed at x0, and the dot denotes the dot product on ℝn.

    Show in article · Read it in context on Wikipedia

    DifferentiableAt.hasGradientAt more general

    theorem DifferentiableAt.hasGradientAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : DifferentiableAt 𝕜 f x → HasGradientAt f (gradient f x) x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 145–147

    Why more general: in the library, having a gradient at a point is by definition the approximation the passage writes with an approximate-equality sign, the error being small beside the displacement

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key c5adcb24bd3e · pair key 55916e79611b.
  59. Theorem · supporting · Relationship with derivative › Relationship with total derivative › Linear approximation to a function

    Not formalized
    This equation is equivalent to the first two terms in the multivariable Taylor series expansion of f at x0.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 00de772c9a85).
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
  60. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Let U be an open set in Rn. If the function f : U → R is differentiable, then the differential of f is the Fréchet derivative of f.

    Show in article · Read it in context on Wikipedia

    hasGradientAt_iff_hasFDerivAt more general

    theorem hasGradientAt_iff_hasFDerivAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x : F} : HasGradientAt f f' x ↔ HasFDerivAt f ((InnerProductSpace.toDual 𝕜 F) f') x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 103–105

    Why more general: the library's fderiv is the Fréchet derivative, and the declaration states an equivalence where the passage states a single direction

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key d9ff3d9d655c · pair key 60378aaba824.

    toDual_gradient more general

    theorem toDual_gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : (InnerProductSpace.toDual 𝕜 F) (gradient f x) = fderiv 𝕜 f x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 128–129

    Why more general: both sides are unconditional in the library, the gradient being set to zero where no derivative exists, and for the passage's differentiable f that case does not arise

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 031e71dae319 · pair key 4266497572b0.
  61. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    Thus ∇f is a function from U to the space Rn such that limh → 0 |f(x + h) − f(x) − ∇f(x) ⋅ h|‖h‖ = 0, where · is the dot product.

    Show in article · Read it in context on Wikipedia

    DifferentiableAt.hasGradientAt more general

    theorem DifferentiableAt.hasGradientAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : DifferentiableAt 𝕜 f x → HasGradientAt f (gradient f x) x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 145–147

    Why more general: Mathlib packs the limit into HasGradientAt, whose unfolding is the same quotient tending to zero written with a variable point approaching x instead of an increment h, and it is stated over any complete inner product space over an RCLike field rather than over real n-space alone.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key df16d1010f8c · pair key c4f8e0ffb553.
  62. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Partial

    Partial: TauCeti states part of this (proved).

    Linearity
    The gradient is linear in the sense that if f and g are two real-valued functions differentiable at the point a ∈ Rn, and α and β are two constants, then αf + βg is differentiable at a, and moreover ∇(αf + βg)(a) = α∇f(a) + β∇g(a).

    Show in article · Read it in context on Wikipedia

    TauCeti.gradient_add states part

    theorem TauCeti.gradient_add {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f g : F → 𝕜} {x : F} : DifferentiableAt 𝕜 f x → DifferentiableAt 𝕜 g x → gradient (f + g) x = gradient f x + gradient g x

    TauCeti.Analysis.Calculus.Gradient · lines 76–79

    Why only part: This is the second conclusion with both coefficients equal to one; the declaration says nothing about scaling either summand.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key ec03f086c334 · pair key 17d75ecd5266.

    DifferentiableAt.add states part

    theorem DifferentiableAt.add {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {F : Type*} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {f g : E → F} {x : E} : DifferentiableAt 𝕜 f x → DifferentiableAt 𝕜 g x → DifferentiableAt 𝕜 (f + g) x

    Mathlib.Analysis.Calculus.FDeriv.Add · lines 211–213

    Why only part: This is the first conclusion with both coefficients equal to one, since the declaration carries no scalar factors.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 8c174dfc456f · pair key b8d2cfe41d79.

    TauCeti.gradient_const_smul states part

    theorem TauCeti.gradient_const_smul {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} (c : 𝕜) : gradient (c • f) x = (starRingEnd 𝕜) c • gradient f x

    TauCeti.Analysis.Calculus.Gradient · lines 83–87

    Why only part: This is the second conclusion with the other coefficient set to zero, so only one term of the combination is covered.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key d721752684a2 · pair key c780f4c751c4.

    DifferentiableAt.const_smul states part

    theorem DifferentiableAt.const_smul {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {F : Type*} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {f : E → F} {x : E} {R : Type*} [Monoid R] [DistribMulAction R F] [SMulCommClass 𝕜 R F] [ContinuousConstSMul R F] : DifferentiableAt 𝕜 f x → ∀ (c : R), DifferentiableAt 𝕜 (c • f) x

    Mathlib.Analysis.Calculus.FDeriv.Add · lines 75–77

    Why only part: This is the first conclusion with the second coefficient set to zero, covering only one summand of the combination.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 4 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key a03d9e44b844 · pair key 208e648c38d7.
  63. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Formalized

    Formalized by 2 parts together (more general):

    1. fderiv_mul Mathlib (proved)
      • the gradient of the product at a equals the value of the first factor at a times the gradient of the second, plus the value of the second at a times the gradient of the first
    2. DifferentiableAt.mul Mathlib (proved)
      • the product of two real-valued functions differentiable at the point a is differentiable at a
    Product rule
    If f and g are real-valued functions differentiable at a point a ∈ Rn, then the product rule asserts that the product fg is differentiable at a, and ∇(fg)(a) = f(a)∇g(a) + g(a)∇f(a).

    Show in article · Read it in context on Wikipedia

    fderiv_mul states part

    theorem fderiv_mul {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {x : E} {𝔸' : Type*} [NormedCommRing 𝔸'] [NormedAlgebra 𝕜 𝔸'] {c d : E → 𝔸'} : DifferentiableAt 𝕜 c x → DifferentiableAt 𝕜 d x → fderiv 𝕜 (c * d) x = c x • fderiv 𝕜 d x + d x • fderiv 𝕜 c x

    Mathlib.Analysis.Calculus.FDeriv.Mul · lines 266–268

    Why only part: Mathlib writes the identity for the differential as a linear functional and over any commutative normed algebra; the gradient is that functional's Riesz vector, so the two sides correspond term by term.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 0042afccb90c · pair key 33eefc917715.

    DifferentiableAt.mul states part

    theorem DifferentiableAt.mul {𝕜 : Type*} [NontriviallyNormedField 𝕜] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {x : E} {𝔸 : Type*} [NormedRing 𝔸] [NormedAlgebra 𝕜 𝔸] {a b : E → 𝔸} : DifferentiableAt 𝕜 a x → DifferentiableAt 𝕜 b x → DifferentiableAt 𝕜 (a * b) x

    Mathlib.Analysis.Calculus.FDeriv.Mul · lines 218–220

    Why only part: The declaration is stated for functions into any normed algebra over any nontrivially normed field, of which real-valued functions on real n-space are one instance.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 296797c29483 · pair key c3afff4be775.
  64. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    First, suppose that the function g is a parametric curve; that is, a function g : I → Rn maps a subset I ⊂ R into Rn. If g is differentiable at a point c ∈ I such that g(c) = a, then (f ∘ g)′(c) = ∇f(a) ⋅ g′(c), where ∘ is the composition operator: (f ∘ g)(x) = f(g(x)).

    Show in article · Read it in context on Wikipedia

    HasFDerivAt.comp_hasDerivAt more general

    theorem HasFDerivAt.comp_hasDerivAt {𝕜 : Type u} [NontriviallyNormedField 𝕜] {F : Type v} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {E : Type w} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {f : 𝕜 → F} {f' : F} (x : 𝕜) {l : F → E} {l' : F →L[𝕜] E} : HasFDerivAt l l' (f x) → HasDerivAt f f' x → HasDerivAt (l ∘ f) (l' f') x

    Mathlib.Analysis.Calculus.Deriv.Comp · lines 375–377

    Why more general: Mathlib keeps the outer derivative as a continuous linear map and the dot product with the gradient is that same map in its Riesz form, and it takes the curve to be defined on the whole line rather than on a subset of it.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key c71d7a6a88c8 · pair key c58ca971e381.
  65. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Not formalized
    More generally, if instead I ⊂ Rk, then the following holds: ∇(f ∘ g)(c) = (Dg(c))T(∇f(a)), where (Dg)T denotes the transpose Jacobian matrix.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 4 candidates examined (receipt a90331227226).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  66. Theorem · main statement · Relationship with derivative › Relationship with Fréchet derivative

    Formalized

    Formalized: Mathlib states a more general result (proved).

    For the second form of the chain rule, suppose that h : I → R is a real valued function on a subset I of R, and that h is differentiable at the point f(a) ∈ I. Then ∇(h ∘ f)(a) = h′(f(a))∇f(a).

    Show in article · Read it in context on Wikipedia

    HasDerivAt.comp_hasFDerivAt more general

    theorem HasDerivAt.comp_hasFDerivAt {𝕜 : Type u} [NontriviallyNormedField 𝕜] {E : Type w} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {𝕜' : Type*} [NontriviallyNormedField 𝕜'] [NormedAlgebra 𝕜 𝕜'] {h₂ : 𝕜' → 𝕜'} {h₂' : 𝕜'} {f : E → 𝕜'} {f' : E →L[𝕜] 𝕜'} (x : E) : HasDerivAt h₂ h₂' (f x) → HasFDerivAt f f' x → HasFDerivAt (h₂ ∘ f) (h₂' • f') x

    Mathlib.Analysis.Calculus.Deriv.Comp · lines 172–174

    Why more general: Mathlib states the outer function on the whole line where the passage restricts it to a subset containing the value of f, and the scalar multiple is taken on the differential rather than on the gradient vector it represents.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key fa9b6e93568e · pair key b4324e7f9de7.
  67. Definition · main statement · Further properties and applications › Level sets

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    A level surface, or isosurface, is the set of all points where some function has a given value.

    Show in article · Read it in context on Wikipedia

    Set.preimage more general

    def Set.preimage {α : Type u} {β : Type v} : (α → β) → Set β → Set α

    Mathlib.Data.Set.Operations · line 134

    Why more general: No library of the universe names a level surface or an isosurface; Mathlib defines the preimage of an arbitrary set, and the passage's object is its value at a one-point set, the extension to larger sets being genuinely more general.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 5b2c8ce04125 · pair key 1cbe326a036f.
  68. Theorem · supporting · Further properties and applications › Level sets

    Formalized

    Formalized: Mathlib states a more general result (proved).

    If f is differentiable, then the dot product (∇f )x ⋅ v of the gradient at a point x with a vector v gives the directional derivative of f at x in the direction v.

    Show in article · Read it in context on Wikipedia

    DifferentiableAt.lineDeriv_eq_fderiv more general

    theorem DifferentiableAt.lineDeriv_eq_fderiv {𝕜 : Type*} [NontriviallyNormedField 𝕜] {F : Type*} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {E : Type*} [NormedAddCommGroup E] [NormedSpace 𝕜 E] {f : E → F} {x v : E} : DifferentiableAt 𝕜 f x → lineDeriv 𝕜 f x v = (fderiv 𝕜 f x) v

    Mathlib.Analysis.Calculus.LineDeriv.Basic · lines 319–321

    Why more general: Mathlib writes the directional derivative as the line derivative and the right-hand side as the differential; the gradient is the vector representing that differential through the inner product, so the two readings give the same number.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 92ddcd56abcf · pair key b8285f6cb416.
  69. Theorem · main statement · Further properties and applications › Level sets

    Formalized

    Formalized: ATLAS (atlas-lean) states a more general result (proved).

    It follows that in this case the gradient of f is orthogonal to the level sets of f.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.levelSet_tangent_vel_of_curve more general

    theorem MetaMathlibExt.levelSet_tangent_vel_of_curve {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] {f : E → ℝ} {c : ℝ} {x v : E} : DifferentiableAt ℝ f x → ∀ {γ : ℝ → E}, γ 0 = x → HasDerivAt γ v 0 → (∀ (t : ℝ), f (γ t) = c) → (fderiv ℝ f x) v = 0

    MathlibExt.Analysis.Calculus.TangentSpaceLevelSet · line 47

    Why more general: The library writes the Frechet derivative where the article writes the gradient, the two being linked by Riesz duality in an inner product space, and it works over an arbitrary real normed space rather than only coordinate space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key dd0d1bfc0147 · pair key 7533cbde0e07.

    TauCeti.span_tangentConeAt_preimage_zero special case

    theorem TauCeti.span_tangentConeAt_preimage_zero {𝕜 : Type*} {E : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup E] [NormedSpace 𝕜 E] [FiniteDimensional 𝕜 E] [NormedAddCommGroup F] [NormedSpace 𝕜 F] {n : WithTop ℕ∞} : n ≠ 0 → ∀ {g : E → F} {g' : E →L[𝕜] F} {y : E}, HasStrictFDerivAt g g' y → (↑g').range = ⊤ → g y = 0 → ∀ {U : Set E}, IsOpen U → y ∈ U → (∀ z ∈ U, ContDiffAt 𝕜 n g z) → Submodule.span 𝕜 (tangentConeAt 𝕜 (g ⁻¹' {0}) y) = (↑g').ker

    TauCeti.Analysis.Calculus.TransverseIntersection · lines 158–179

    Why a special case: It assumes a surjective strict derivative and local smoothness, so it covers the regular points among the article's differentiable functions, and it also supplies the reverse inclusion of the tangent space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 6aeead9788d8 · pair key d61962758597.
  70. Theorem · supporting · Further properties and applications › Level sets

    Formalized

    Formalized: ATLAS (atlas-lean) states a more general result (proved).

    For example, a level surface in three-dimensional space is defined by an equation of the form F(x, y, z) = c. The gradient of F is then normal to the surface.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.levelSet_tangent_vel_of_curve more general

    theorem MetaMathlibExt.levelSet_tangent_vel_of_curve {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] {f : E → ℝ} {c : ℝ} {x v : E} : DifferentiableAt ℝ f x → ∀ {γ : ℝ → E}, γ 0 = x → HasDerivAt γ v 0 → (∀ (t : ℝ), f (γ t) = c) → (fderiv ℝ f x) v = 0

    MathlibExt.Analysis.Calculus.TangentSpaceLevelSet · line 47

    Why more general: The declaration is stated for an arbitrary real normed space and an arbitrary level value, so the article's three dimensional surface is one instance of it, and it uses the Frechet derivative in place of the gradient vector.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key abc1fcbbec25 · pair key 84f86db8e0a7.
  71. Theorem · main statement · Further properties and applications › Level sets

    Not formalized
    More generally, any embedded hypersurface in a Riemannian manifold can be cut out by an equation of the form F(P) = 0 such that dF is nowhere zero.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt 780815c94216).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  72. Theorem · main statement · Further properties and applications › Level sets

    Partial

    Partial: ATLAS (atlas-lean) states a special case (proved).

    The gradient of F is then normal to the hypersurface.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.levelSet_tangent_vel_of_curve special case

    theorem MetaMathlibExt.levelSet_tangent_vel_of_curve {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] {f : E → ℝ} {c : ℝ} {x v : E} : DifferentiableAt ℝ f x → ∀ {γ : ℝ → E}, γ 0 = x → HasDerivAt γ v 0 → (∀ (t : ℝ), f (γ t) = c) → (fderiv ℝ f x) v = 0

    MathlibExt.Analysis.Calculus.TangentSpaceLevelSet · line 47

    Why a special case: Its ambient space is a flat Euclidean one rather than a general Riemannian manifold, so it covers the flat instances of the passage, and it needs no regularity of the differential.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 698f4d42bc78 · pair key 200b45afa9d2.

    TauCeti.span_tangentConeAt_preimage_zero special case

    theorem TauCeti.span_tangentConeAt_preimage_zero {𝕜 : Type*} {E : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup E] [NormedSpace 𝕜 E] [FiniteDimensional 𝕜 E] [NormedAddCommGroup F] [NormedSpace 𝕜 F] {n : WithTop ℕ∞} : n ≠ 0 → ∀ {g : E → F} {g' : E →L[𝕜] F} {y : E}, HasStrictFDerivAt g g' y → (↑g').range = ⊤ → g y = 0 → ∀ {U : Set E}, IsOpen U → y ∈ U → (∀ z ∈ U, ContDiffAt 𝕜 n g z) → Submodule.span 𝕜 (tangentConeAt 𝕜 (g ⁻¹' {0}) y) = (↑g').ker

    TauCeti.Analysis.Calculus.TransverseIntersection · lines 158–179

    Why a special case: Its ambient space is flat and finite dimensional and it asks for a strict derivative, so it is the passage restricted to hypersurfaces in Euclidean space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 21ffc13069e3 · pair key 61779ea45aa1.
  73. Definition · supporting · Further properties and applications › Level sets

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    Similarly, an affine algebraic hypersurface may be defined by an equation F(x1, ..., xn) = 0, where F is a polynomial.

    Show in article · Read it in context on Wikipedia

    MvPolynomial.zeroLocus more general

    def MvPolynomial.zeroLocus {k : Type*} (K : Type*) [Field k] [Field K] [Algebra k K] {σ : Type*} : Ideal (MvPolynomial σ k) → Set (σ → K)

    Mathlib.RingTheory.Nullstellensatz · lines 39–40

    Why more general: Mathlib defines the common zero set of a whole ideal of polynomials and lets the coordinates live in an extension field, which covers arbitrary affine algebraic sets; the article's hypersurface is the case of a single generator.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 2c981e9a8058 · pair key f04d76b2a7de.
  74. Definition · main statement · Further properties and applications › Level sets

    Partial

    Partial: ATLAS (atlas-lean) states a special case (a Prop-valued definition: it states the claim without proving it).

    The gradient of F is zero at a singular point of the hypersurface (this is the definition of a singular point).

    Show in article · Read it in context on Wikipedia

    AlgebraicGeometry.CubicSurface.IsSingularPoint special case

    def AlgebraicGeometry.CubicSurface.IsSingularPoint {k : Type*} [Field k] : AlgebraicGeometry.CubicSurface k → (Fin 4 → k) → Prop

    MathlibExt.AlgebraicGeometry.CubicSurface · line 92

    Why a special case: Its polynomial is restricted to homogeneous cubics in four variables, so it is the article's notion of singular point for that family of affine hypersurfaces.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 2c04396fb8a1 · pair key 1d1fd76d5d23.

    WeierstrassCurve.Affine.IsSingular special case

    def WeierstrassCurve.Affine.IsSingular {R : Type*} [CommRing R] : WeierstrassCurve.Affine R → R → R → Prop

    TauCeti.AlgebraicGeometry.EllipticCurve.Affine.Singular · lines 95–96

    Why a special case: Its hypersurface is a plane curve given by a Weierstrass polynomial in two variables, so it is the article's definition for that family.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 875f47aea0c3 · pair key bfc483c81321.

    AlgebraicGeometry.CubicSurface.isSingularPoint_iff special case

    theorem AlgebraicGeometry.CubicSurface.isSingularPoint_iff {k : Type*} [Field k] : 3 ≠ 0 → ∀ (S : AlgebraicGeometry.CubicSurface k) (x : Fin 4 → k), S.IsSingularPoint x ↔ ∀ (i : Fin 4), (MvPolynomial.aeval x) ((MvPolynomial.pderiv i) S.defining) = 0

    MathlibExt.AlgebraicGeometry.CubicSurface · line 116

    Why a special case: The equivalence is proved for cubic surfaces only, and Euler's identity lets it drop the requirement that the point lie on the hypersurface.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 89fea632baad · pair key eefdd6a4b1da.
  75. Theorem · main statement · Further properties and applications › Level sets

    Formalized

    Formalized: TauCeti states a more general result (proved).

    At a non-singular point, it is a nonzero normal vector.

    Show in article · Read it in context on Wikipedia

    TauCeti.span_tangentConeAt_preimage_zero more general

    theorem TauCeti.span_tangentConeAt_preimage_zero {𝕜 : Type*} {E : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup E] [NormedSpace 𝕜 E] [FiniteDimensional 𝕜 E] [NormedAddCommGroup F] [NormedSpace 𝕜 F] {n : WithTop ℕ∞} : n ≠ 0 → ∀ {g : E → F} {g' : E →L[𝕜] F} {y : E}, HasStrictFDerivAt g g' y → (↑g').range = ⊤ → g y = 0 → ∀ {U : Set E}, IsOpen U → y ∈ U → (∀ z ∈ U, ContDiffAt 𝕜 n g z) → Submodule.span 𝕜 (tangentConeAt 𝕜 (g ⁻¹' {0}) y) = (↑g').ker

    TauCeti.Analysis.Calculus.TransverseIntersection · lines 158–179

    Why more general: The declaration covers continuously differentiable maps into any normed space over an RCLike field, of which the article's real polynomial is one instance, and it also gives the reverse inclusion of the tangent space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key e6a414468938 · pair key dbe6987d5dfd.

    MetaMathlibExt.levelSet_tangent_vel_of_curve states part

    theorem MetaMathlibExt.levelSet_tangent_vel_of_curve {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] {f : E → ℝ} {c : ℝ} {x v : E} : DifferentiableAt ℝ f x → ∀ {γ : ℝ → E}, γ 0 = x → HasDerivAt γ v 0 → (∀ (t : ℝ), f (γ t) = c) → (fderiv ℝ f x) v = 0

    MathlibExt.Analysis.Calculus.TangentSpaceLevelSet · line 47

    Why only part: It speaks of any point of the level set and addresses only the direction of the gradient, leaving the non-vanishing that the article reads off non-singularity to the definition of a singular point.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key c3878b91d9f1 · pair key 0724e3bbe781.
  76. Definition · main statement · Further properties and applications › Conservative vector fields and the gradient theorem

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    The gradient of a function is called a gradient field.

    Show in article · Read it in context on Wikipedia

    gradient more general

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why more general: Mathlib builds the gradient over any complete inner product space over the reals or complexes, and it returns the zero vector at points where no gradient exists.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 6458090c2725 · pair key 0150b067e7d1.
  77. Theorem · main statement · Further properties and applications › Conservative vector fields and the gradient theorem

    Formalized

    Formalized: TauCeti states a more general result (proved).

    A (continuous) gradient field is always a conservative vector field: its line integral along any path depends only on the endpoints of the path, and can be evaluated by the gradient theorem (the fundamental theorem of calculus for line integrals).

    Show in article · Read it in context on Wikipedia

    AbsolutelyContinuousOnInterval.integral_fderiv_apply_eq_sub more general

    theorem AbsolutelyContinuousOnInterval.integral_fderiv_apply_eq_sub {F : Type*} [NormedAddCommGroup F] [NormedSpace ℝ F] {γ V : ℝ → F} {a b : ℝ} : AbsolutelyContinuousOnInterval γ a b → (∀ᵐ (t : ℝ), t ∈ Set.uIoc a b → HasDerivAt γ (V t) t) → ∀ {φ : F → ℝ}, ContDiff ℝ 1 φ → ∫ (t : ℝ) in a..b, (fderiv ℝ φ (γ t)) (V t) = φ (γ b) - φ (γ a)

    TauCeti.MeasureTheory.Integral.IntervalIntegral.AbsolutelyContinuousFun · lines 60–67

    Why more general: The field is paired with the velocity through the Fréchet differential of the potential rather than through the inner product with the gradient vector; in a Euclidean space these are the same pairing, and the ambient space is left as any real normed space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key cd68a760d2f8 · pair key 6923e85be27f.

    TauCeti.IsIntegralCurveOn.integral_norm_gradient_sq_eq_sub states part

    theorem TauCeti.IsIntegralCurveOn.integral_norm_gradient_sq_eq_sub {E : Type*} [NormedAddCommGroup E] [InnerProductSpace ℝ E] [CompleteSpace E] {f : E → ℝ} {γ : ℝ → E} {s : Set ℝ} {a b : ℝ} : IsIntegralCurveOn γ (fun x x_1 ↦ -gradient f x_1) s → Set.uIcc a b ⊆ s → (∀ t ∈ Set.uIcc a b, DifferentiableAt ℝ f (γ t)) → IntervalIntegrable (fun t ↦ ‖gradient f (γ t)‖ ^ 2) MeasureTheory.MeasureSpace.volume a b → ∫ (t : ℝ) in a..b, ‖gradient f (γ t)‖ ^ 2 = f (γ a) - f (γ b)

    TauCeti.Analysis.Calculus.Morse.GradientFlow · lines 140–158

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 0dc1349c8825 · pair key 499eba15de36.

    intervalIntegral.integral_deriv_eq_sub states part

    theorem intervalIntegral.integral_deriv_eq_sub {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] {a b : ℝ} [CompleteSpace E] {f : ℝ → E} : (∀ x ∈ Set.uIcc a b, DifferentiableAt ℝ f x) → IntervalIntegrable (deriv f) MeasureTheory.MeasureSpace.volume a b → ∫ (y : ℝ) in a..b, deriv f y = f b - f a

    Mathlib.MeasureTheory.Integral.IntervalIntegral.FundThmCalculus · lines 1176–1178

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key c1919ae92c5a · pair key b71a0448e002.
  78. Theorem · main statement · Further properties and applications › Conservative vector fields and the gradient theorem

    Partial

    Partial: Mathlib states a special case (proved).

    Conversely, a (continuous) conservative vector field is always the gradient of a function.

    Show in article · Read it in context on Wikipedia

    Convex.exists_forall_hasFDerivAt_of_fderiv_symmetric special case

    theorem Convex.exists_forall_hasFDerivAt_of_fderiv_symmetric {𝕜 : Type*} {E : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup E] [NormedSpace 𝕜 E] [NormedAddCommGroup F] [NormedSpace 𝕜 F] [NormedSpace ℝ E] [NormedSpace ℝ F] {s : Set E} {ω : E → E →L[𝕜] F} [CompleteSpace F] : Convex ℝ s → IsOpen s → DifferentiableOn ℝ ω s → (∀ a ∈ s, ∀ (x y : E), ((fderiv ℝ ω a) x) y = ((fderiv ℝ ω a) y) x) → ∃ f, ∀ a ∈ s, HasFDerivAt f (ω a) a

    Mathlib.MeasureTheory.Integral.CurveIntegral.Poincare · lines 375–382

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 295eaaadf612 · pair key 95a171e7e353.
  79. Theorem · main statement · Further properties and applications › Gradient is direction of steepest ascent

    Not formalized
    The gradient of a function f: ℝn → ℝ at point x is also the direction of its steepest ascent, i.e. it maximizes its directional derivative:

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 4 queries and 3 candidates examined (receipt c4dd2d9a6943).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  80. Theorem · supporting · Further properties and applications › Gradient is direction of steepest ascent

    Partial

    Partial: Mathlib states part of this (proved).

    Let v ∈ ℝn be an arbitrary unit vector. With the directional derivative defined as ∇vf(x) = limh → 0 f(x + vh) − f(x)h, we get, by substituting the function f(x + vh) with its Taylor series, ∇vf(x) = limh → 0 (f(x) + ∇f ⋅ vh + R) − f(x)h, where R denotes higher order terms in vh.

    Show in article · Read it in context on Wikipedia

    DifferentiableAt.hasGradientAt states part

    theorem DifferentiableAt.hasGradientAt {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x : F} : DifferentiableAt 𝕜 f x → HasGradientAt f (gradient f x) x

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 145–147

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 42e258e4fa66 · pair key b20e33cfba3d.

    hasGradientAt_iff_isLittleO_nhds_zero states part

    theorem hasGradientAt_iff_isLittleO_nhds_zero {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {f' x : F} : HasGradientAt f f' x ↔ (fun h ↦ f (x + h) - f x - Inner.inner 𝕜 f' h) =o[𝓝 0] fun h ↦ h

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 269–271

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 64d4e8b6ea1e · pair key 12b7268988b6.

    hasLineDerivAt_iff_tendsto_slope_zero states part

    theorem hasLineDerivAt_iff_tendsto_slope_zero {𝕜 : Type*} [NontriviallyNormedField 𝕜] {F : Type*} [NormedAddCommGroup F] [NormedSpace 𝕜 F] {E : Type*} [AddCommGroup E] [_root_.Module 𝕜 E] {f : E → F} {f' : F} {x v : E} : HasLineDerivAt 𝕜 f f' x v ↔ Filter.Tendsto (fun t ↦ t⁻¹ • (f (x + t • v) - f x)) (𝓝[≠] 0) (𝓝 f')

    Mathlib.Analysis.Calculus.LineDeriv.Basic · lines 213–217

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 54521513f85b · pair key b9e9ef2595f0.
  81. Theorem · supporting · Further properties and applications › Gradient is direction of steepest ascent

    Partial

    Partial: Mathlib states part of this (proved).

    Dividing by h, and taking the limit yields a term which is bounded from above by the Cauchy–Schwarz inequality[8] |∇vf(x)| = |∇f ⋅ v| ≤ ||∇f||||v|| = ||∇f||.

    Show in article · Read it in context on Wikipedia

    abs_real_inner_le_norm states part

    theorem abs_real_inner_le_norm {F : Type*} [SeminormedAddCommGroup F] [InnerProductSpace ℝ F] (x y : F) : |Inner.inner ℝ x y| ≤ ‖x‖ * ‖y‖

    Mathlib.Analysis.InnerProductSpace.Basic · lines 471–472

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 773a268a681c · pair key 40edcce4c799.

    inner_gradient_left states part

    theorem inner_gradient_left {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] {f : F → 𝕜} {x y : F} : Inner.inner 𝕜 (gradient f x) y = (fderiv 𝕜 f x) y

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 293–294

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 56d3948daee4 · pair key 8418bdba775d.
  82. Theorem · main statement · Further properties and applications › Gradient is direction of steepest ascent

    Not formalized
    Choosing v∗ = ∇f/||∇f|| maximizes the directional derivative, and equals the upper bound |∇v∗f(x)| = ||∇f||2/||∇f|| = ||∇f||.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 5 queries and 4 candidates examined (receipt f820ea00e723).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  83. Definition · main statement · Generalizations › Jacobian

    Formalized

    Formalized: ATLAS (atlas-lean) states this exactly (a definition).

    Suppose f : Rn → Rm is a function such that each of its first-order partial derivatives exist on ℝn. Then the Jacobian matrix of f is defined to be an m×n matrix, denoted by J𝕗(𝕩) or simply J.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.jacobianMatrix exact match

    def MetaMathlibExt.jacobianMatrix {m n : ℕ} : ((Fin m → ℝ) → Fin n → ℝ) → (Fin m → ℝ) → Matrix (Fin n) (Fin m) ℝ

    MathlibExt.Analysis.Calculus.Jacobian · line 39

    Note: Where a first partial fails to exist the Lean derivative returns zero, so the Lean object is total; the passage speaks only of maps whose first partials all exist, and that default plays no role for them.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 55c7dcb59ac7 · pair key e6772f77bca6.
  84. Definition · main statement · Generalizations › Jacobian

    Partial

    Partial: ATLAS (atlas-lean) states this up to a convention difference, but as a proved theorem, not a definition; a definition passage is formalized only by a Lean definition.

    The (i,j)th entry is Jij = ∂fi/∂xj.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.jacobianMatrix_apply close match

    theorem MetaMathlibExt.jacobianMatrix_apply {m n : ℕ} (f : (Fin m → ℝ) → Fin n → ℝ) (x : Fin m → ℝ) (i : Fin n) (j : Fin m) : MetaMathlibExt.jacobianMatrix f x i j = deriv (fun t ↦ f (x + t • Pi.single j 1) i) 0

    MathlibExt.Analysis.Calculus.Jacobian · line 48

    Why close, not exact: It writes the partial derivative of a component with respect to one variable as the one-variable derivative at zero of that component along the corresponding coordinate axis, which is what that partial means.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 0ce90fa710ba · pair key 97fbc06cb12b.
  85. Theorem · supporting · Generalizations › Jacobian

    Formalized

    Formalized: ATLAS (atlas-lean) states this up to a convention difference (proved).

    Explicitly J = [∂f∂x1⋯∂f∂xn] = [∇Tf1⋮∇Tfm] = [∂f1∂x1⋯∂f1∂xn⋮⋱⋮∂fm∂x1⋯∂fm∂xn].

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.jacobianMatrix_apply close match

    theorem MetaMathlibExt.jacobianMatrix_apply {m n : ℕ} (f : (Fin m → ℝ) → Fin n → ℝ) (x : Fin m → ℝ) (i : Fin n) (j : Fin m) : MetaMathlibExt.jacobianMatrix f x i j = deriv (fun t ↦ f (x + t • Pi.single j 1) i) 0

    MathlibExt.Analysis.Calculus.Jacobian · line 48

    Why close, not exact: The library spells the (i,j) entry out as the derivative at 0 of the one-variable restriction of component i along the j-th coordinate axis, which is what the passage's partial-derivative symbol denotes; columns and rows of that matrix are then the per-variable partials and the transposed component gradients the passage lists.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key b18a7664248c · pair key 0098574d3a01.

    MetaMathlibExt.jacobianMatrix_apply_of_differentiableAt special case

    theorem MetaMathlibExt.jacobianMatrix_apply_of_differentiableAt {m n : ℕ} (f : (Fin m → ℝ) → Fin n → ℝ) (x : Fin m → ℝ) : DifferentiableAt ℝ f x → ∀ (i : Fin n) (j : Fin m), MetaMathlibExt.jacobianMatrix f x i j = (fderiv ℝ f x) (Pi.single j 1) i

    MathlibExt.Analysis.Calculus.Jacobian · line 67

    Why a special case: It assumes full Fréchet differentiability at the point, where the passage assumes only that the first-order partials exist, so it covers a sub-family of the passage's maps.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Not re-checked: sampled out of challenge run chal-tierA-b1 by jack, in which Claude Opus 5 (claude-opus-5) re-checked a sample on 1 Nov 2026 (UTC).
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 3270046cf1e8 · pair key 148573201af3.
  86. Theorem · supporting · Generalizations › Gradient of a vector field

    Formalized

    Formalized: Mathlib states a more general result (a definition).

    Since the total derivative of a vector field is a linear mapping from vectors to vectors, it is a tensor quantity.

    Show in article · Read it in context on Wikipedia

    fderiv more general

    def fderiv (𝕜 : Type*) [NontriviallyNormedField 𝕜] {E : Type*} [AddCommGroup E] [_root_.Module 𝕜 E] [TopologicalSpace E] {F : Type*} [AddCommGroup F] [_root_.Module 𝕜 F] [TopologicalSpace F] : (E → F) → E → E →L[𝕜] F

    Mathlib.Analysis.Calculus.FDeriv.Defs · lines 160–161

    Why more general: Lean carries the assertion in the type: the total derivative is produced as a continuous linear map between the two spaces, so linearity is part of the object rather than a separate proposition, and the scalars and spaces are left arbitrary.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 818c032d3d53 · pair key ff61c9a335f4.
  87. Theorem · supporting · Generalizations › Gradient of a vector field

    Not formalized
    Since the total derivative of a vector field is a linear mapping from vectors to vectors, it is a tensor quantity.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 20a384dcb552).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  88. Definition · main statement · Generalizations › Gradient of a vector field

    Not formalized
    In rectangular coordinates, the gradient of a vector field f = ( f1, f2, f3) is defined by: ∇f = gjk∂fi∂xjei ⊗ ek, (where the Einstein summation notation is used and the tensor product of the vectors ei and ek is a dyadic tensor of type (2,0)).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 4555aebca283).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  89. Theorem · main statement · Generalizations › Gradient of a vector field

    Not formalized
    Overall, this expression equals the transpose of the Jacobian matrix: ∂fi∂xj = ∂(f1, f2, f3)∂(x1, x2, x3).

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 49914210b866).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  90. Theorem · main statement · Generalizations › Gradient of a vector field

    Not formalized
    In curvilinear coordinates, or more generally on a curved manifold, the gradient involves Christoffel symbols: ∇f = gjk(∂fi∂xj + Γijlfl)ei ⊗ ek, where gjk are the components of the inverse metric tensor and the ei are the coordinate basis vectors.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 8a915dedd432).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  91. Definition · main statement · Generalizations › Gradient of a vector field

    Not formalized
    Expressed more invariantly, the gradient of a vector field f can be defined by the Levi-Civita connection and metric tensor:[11] ∇afb = gac∇cfb, where ∇c is the connection.

    Show in article · Read it in context on Wikipedia

    Searched 48 libraries — not found. Which libraries, at which commits (pinned 2026-10-17).

    Details
    • Searched by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: the search receipt lists 3 queries and 2 candidates examined (receipt 50056c4911ef).
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: upheld.
  92. Definition · main statement · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states this exactly (a definition).

    For any smooth function f on a Riemannian manifold (M, g), the gradient of f is the vector field ∇f such that for any vector field X, g(∇f, X) = ∂Xf, that is, gx((∇f)x, Xx) = (∂Xf)(x), where gx( , ) denotes the inner product of tangent vectors at x defined by the metric g and ∂X f is the function that takes any point x ∈ M to the directional derivative of f in the direction X, evaluated at x.

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient exact match

    def TauCeti.Manifold.mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] (I : ModelWithCorners ℝ E H) {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] : (M → ℝ) → (x : M) → TangentSpace I x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 59–60

    Note: Lean returns the zero tangent vector where the function has no differential, a default value the passage's smooth functions never meet, and the pairing used is the inner product the metric puts on each tangent space.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 8e34e3ea76f2 · pair key 27b0033b43e6.

    TauCeti.Manifold.inner_mgradient close match

    theorem TauCeti.Manifold.inner_mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] {f : M → ℝ} {x : M} (v : TangentSpace I x) : Inner.inner ℝ (TauCeti.Manifold.mgradient I f x) v = (d% f x) v

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 71–73

    Why close, not exact: The library states the defining identity pointwise: the metric inner product of the gradient at a point with a tangent vector is the differential of the function applied to that vector, with the directional derivative written as that differential.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 0e97f06d2244 · pair key 6c79b20d9f08.

    TauCeti.Manifold.eq_mgradient_iff more general

    theorem TauCeti.Manifold.eq_mgradient_iff {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] {f : M → ℝ} {x : M} {w : TangentSpace I x} : w = TauCeti.Manifold.mgradient I f x ↔ ∀ (v : TangentSpace I x), Inner.inner ℝ w v = (d% f x) v

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 83–85

    Why more general: It states the full characterisation, that a tangent vector is the gradient exactly when its metric pairing reproduces the differential, so it also settles uniqueness of the vector field the passage names.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 3 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key f70d3aacc379 · pair key 13cf1ae18a69.

    gradient special case

    def gradient {𝕜 : Type*} {F : Type*} [RCLike 𝕜] [NormedAddCommGroup F] [InnerProductSpace 𝕜 F] [CompleteSpace F] : (F → 𝕜) → F → F

    Mathlib.Analysis.Calculus.Gradient.Basic · lines 83–84

    Why a special case: It covers the flat case only, an inner product space with its constant metric rather than a Riemannian manifold with a varying one.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 4 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 3dfb8e2.
    • Location key 470fdba3430e · pair key 35502d6d91cf.
  93. Definition · supporting · Generalizations › Riemannian manifolds

    Partial

    Partial: ATLAS (atlas-lean) states a special case (proved).

    In other words, in a coordinate chart φ from an open subset of M to an open subset of Rn, (∂X f )(x) is given by: ∑nj = 1 Xj(φ(x))∂∂xj(f ∘ φ−1)|φ(x), where Xj denotes the jth component of X in this coordinate chart.

    Show in article · Read it in context on Wikipedia

    MetaMathlibExt.partialDeriv_eq_fderiv_apply special case

    theorem MetaMathlibExt.partialDeriv_eq_fderiv_apply {n : ℕ} (f : (Fin n → ℝ) → ℝ) (i : Fin n) (x : Fin n → ℝ) : DifferentiableAt ℝ f x → MetaMathlibExt.partialDeriv f i x = (fderiv ℝ f x) (Pi.single i 1)

    MathlibExt.Analysis.Calculus.HigherPartials · line 66

    Why a special case: The declaration mentions neither a chart nor a vector field: at the identity chart and an axis direction the article's sum over coordinates has a single surviving term.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 64c2773.
    • Location key 47cb638d35fa · pair key 1d52025bb9dc.
  94. Theorem · main statement · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states a more general result (proved).

    So, the local form of the gradient takes the form: ∇f = gik∂f∂xkei.

    Show in article · Read it in context on Wikipedia

    Riemannian.Tensor.rieszDual_eq_sum_chartLocalFrame more general

    theorem Riemannian.Tensor.rieszDual_eq_sum_chartLocalFrame {E : Type*} […] {H : Type*} {I : ModelWithCorners ℝ E H} {M : Type*} (a : M) {x : M} : x ∈ (chartAt H a).source → ∀ (φ : TangentSpace I x →L[ℝ] ℝ), (Riemannian.Tensor.rieszDual x) φ = ∑ i, (Riemannian.Tensor.chartInvGramMatrix a x).mulVec (fun j ↦ φ (Riemannian.Tensor.chartLocalFrame a j x)) i • Riemannian.Tensor.chartLocalFrame a i x
    Full type, as Lean prints it
    ∀ {E : Type u_1} [inst : NormedAddCommGroup E] [inst_1 : NormedSpace ℝ E] [inst_2 : FiniteDimensional ℝ E] {H : Type u_2} [inst_3 : TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type u_3} [inst_4 : TopologicalSpace M] [inst_5 : ChartedSpace H M] [inst_6 : IsManifold I 1 M] [inst_7 : Bundle.RiemannianBundle fun x ↦ TangentSpace I x] (a : M) {x : M}, x ∈ (chartAt H a).source → ∀ (φ : TangentSpace I x →L[ℝ] ℝ), (Riemannian.Tensor.rieszDual x) φ = ∑ i, (Riemannian.Tensor.chartInvGramMatrix a x).mulVec (fun j ↦ φ (Riemannian.Tensor.chartLocalFrame a j x)) i • Riemannian.Tensor.chartLocalFrame a i x

    TauCeti.Geometry.Manifold.VectorBundle.Riemannian.Riesz · lines 99–136

    Why more general: It is stated for an arbitrary covector on the tangent fibre, of which the differential of f is one; its inverse Gram matrix is the article's inverse metric components and its frame is the chart trivialization of a fixed basis of the model space, that is the chart's coordinate directions.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key b1d644804770 · pair key 200c8d3b794f.
  95. Theorem · supporting · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states a more general result (proved).

    Generalizing the case M = Rn, the gradient of a function is related to its exterior derivative, since (∂Xf)(x) = (df)x(Xx). More precisely, the gradient ∇f is the vector field associated to the differential 1-form df using the musical isomorphism ♯ = ♯g: T∗M → TM (called "sharp")…

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.inner_mgradient more general

    theorem TauCeti.Manifold.inner_mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] {f : M → ℝ} {x : M} (v : TangentSpace I x) : Inner.inner ℝ (TauCeti.Manifold.mgradient I f x) v = (d% f x) v

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 71–73

    Why more general: It makes determinate what the sentence announces in loose terms: pairing the gradient with a tangent vector under the metric returns the value of the differential of f on that vector.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key b86aeb4a7cb2 · pair key 6e76e48d9050.

    TauCeti.Manifold.mgradient_def more general

    theorem TauCeti.Manifold.mgradient_def {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] (f : M → ℝ) (x : M) : TauCeti.Manifold.mgradient I f x = (Riemannian.Tensor.rieszDual x) (d% f x)

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 63–65

    Why more general: It states the link in the other direction from the first card, giving the gradient itself in terms of the differential rather than pairing it with a vector.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key fdb64ba0cfc0 · pair key 60e550a56d74.
  96. Theorem · supporting · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states a more general result (proved).

    Generalizing the case M = Rn, the gradient of a function is related to its exterior derivative, since (∂Xf)(x) = (df)x(Xx). More precisely, the gradient ∇f is the vector field associated to the differential 1-form df using the musical isomorphism ♯ = ♯g: T∗M → TM (called "sharp") defined by the metric g.

    Show in article · Read it in context on Wikipedia

    TauCeti.mextDeriv_ofFunction more general

    theorem TauCeti.mextDeriv_ofFunction {E : Type*} {H : Type*} {M : Type*} {F : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [TopologicalSpace H] {I : ModelWithCorners ℝ E H} [TopologicalSpace M] [ChartedSpace H M] [NormedAddCommGroup F] [NormedSpace ℝ F] (f : M → F) (x : M) : TauCeti.mextDeriv (TauCeti.RoughForm.ofFunction I f) x = (ContinuousAlternatingMap.ofSubsingleton ℝ (TangentSpace I x) F 0) (d% f x)

    TauCeti.Geometry.Manifold.DifferentialForm.ExtDeriv · lines 345–350

    Why more general: It is stated for a function into any real normed space, and it packages the derivative as an alternating map in one argument, which is the one-form the article writes.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 48138f649fcc · pair key 8aba974d4335.
  97. Definition · main statement · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states this exactly (a definition).

    Generalizing the case M = Rn, the gradient of a function is related to its exterior derivative, since (∂Xf)(x) = (df)x(Xx). More precisely, the gradient ∇f is the vector field associated to the differential 1-form df using the musical isomorphism ♯ = ♯g: T∗M → TM (called "sharp") defined by the metric g.

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient exact match

    def TauCeti.Manifold.mgradient {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] (I : ModelWithCorners ℝ E H) {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] : (M → ℝ) → (x : M) → TangentSpace I x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 59–60

    Note: The map it uses is the fibrewise Riesz equivalence from the dual of a tangent space onto that tangent space determined by the metric, which is the raising map the article names; the only value it supplies outside the article's reach is the zero derivative of a function with no derivative, which carries no content.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key d7bd143b4778 · pair key 6223f90480de.

    TauCeti.Manifold.mgradient_def close match

    theorem TauCeti.Manifold.mgradient_def {E : Type*} [NormedAddCommGroup E] [NormedSpace ℝ E] [FiniteDimensional ℝ E] {H : Type*} [TopologicalSpace H] {I : ModelWithCorners ℝ E H} {M : Type*} [TopologicalSpace M] [ChartedSpace H M] [IsManifold I 1 M] [Bundle.RiemannianBundle fun x ↦ TangentSpace I x] (f : M → ℝ) (x : M) : TauCeti.Manifold.mgradient I f x = (Riemannian.Tensor.rieszDual x) (d% f x)

    TauCeti.Geometry.Manifold.Riemannian.Gradient.Basic · lines 63–65

    Why close, not exact: It unfolds the library's gradient into the defining expression the passage uses, naming the fibrewise metric dual of the derivative of f.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 2 for this passage, high confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 183142898db4 · pair key d68b6a0022b7.
  98. Theorem · supporting · Generalizations › Riemannian manifolds

    Formalized

    Formalized: TauCeti states a more general result (proved).

    The relation between the exterior derivative and the gradient of a function on Rn is a special case of this in which the metric is the flat metric given by the dot product.

    Show in article · Read it in context on Wikipedia

    TauCeti.Manifold.mgradient_eq_gradient more general

    theorem TauCeti.Manifold.mgradient_eq_gradient {F : Type*} [NormedAddCommGroup F] [InnerProductSpace ℝ F] [FiniteDimensional ℝ F] (f : F → ℝ) (x : F) : have x_1 := { g := (riemannianMetricVectorSpace F).toRiemannianMetric }; TauCeti.Manifold.mgradient (modelWithCornersSelf ℝ F) f x = gradient f x

    TauCeti.Geometry.Manifold.Riemannian.Gradient.NormedSpace · lines 44–61

    Why more general: It is stated for an arbitrary finite-dimensional real inner product space with the metric built from its inner product, and the article's coordinate space with the dot product is one such space; this is the flat case of the manifold statement, as the sentence says.

    Details
    • Found by Claude Opus 5 (claude-opus-5) on 31 Oct–1 Nov 2026 (UTC), locate run loc-tierA-b1b by jack: candidate 1 for this passage, medium confidence.
    • Re-checked by Claude Opus 5 (claude-opus-5) on 1 Nov 2026 (UTC), challenge run chal-tierA-b1 by jack: it survived.
    • Record set tierA-b1b, exported 1 Nov 2026 by formal-companion d85385a.
    • Lean code: the declaration's type as Lean prints it, laid out as a header (the corpus keeps no hypothesis names), from the declaration corpus of universe u1 at commit 83245ab.
    • Location key 2e22185610e6 · pair key de06ee815a45.

Data: view model (JSON) · Literature Links records (JSONL)