Engine

Scientific audit layer

SciCoQA taxonomy, FaithformBench silent-fix gate, synthetic mutations. Auditor outcomes are satisfies, violates, inconsistent, or unknown — never a forced true/false.

Claim → contract → evidence → verdict. Each stage upgrades independently. This page is the SciCoQA taxonomy plus the FaithformBench gate. It is not eight agents in one repo.

satisfies

1

violates

9

inconsistent

6

unknown

2

Paper↔code codes

  • Parameter mismatch2
  • Equation mismatch8
  • Missing operation2
  • Extra operation0
  • Default-value mismatch2
  • Constraint mismatch6
  • Shape mismatch0
  • Algorithm mismatch3
  • Training mismatch2
  • Evaluation mismatch5
  • Omitted implementation detail3
  • Dependency/version mismatch0

Faithfulness gate

A formalizer that changes σ ∈ (−1, 1) to {±1} will prove Lorentz membership of a theorem the paper did not state. F and P are separate.

FormalizationFPAccept?

σ ∈ (−1, 1)

σ ∈ (−1, 1)

Keeps the paper. Isometry is false on the whole domain.

truefalsefails

σ ∈ (−1, 1)

σ ∈ {−1, +1}

Sycophantic repair. Lean would prove O-membership of a theorem the paper did not state.

falsetruefails

Synthetic mutations

Start from a correct Lorentz operator, mutate as SciCoQA does, require an oracle to fire. 6/6 caught.

  • equation⟨·,·⟩_LEuclidean dotholds
  • parameterσᵣ per relationone flip_sign per moduleholds
  • constraintσ ∈ {±1}σ ∈ (−1, 1)holds
  • missingR BR onlyholds
  • constraintσ ∈ (−1, 1) as printedσ ∈ {±1} after ‘helpfulness’holds
  • algorithmLorentz projectionv = d + (c·d)cholds

Upgrade order

  1. 1. Claim graph [Atlas]
  2. 2. Paper↔code taxonomy [SciCoQA]
  3. 3. Faithfulness gate [FaithformBench]
  4. 4. Contracts (not silent-fixed) [FormalScience]
  5. 5. Proof DAG [LeanMarathon]
  6. 6. Repository verification [Vero]
  7. 7. Scientific execution [SWE-bench Science]
  8. 8. Experiment reproduction [PaperBench]

Lean is not in this build. The gate exists so a future autoformalizer cannot silently repair FlorE. Invariants.