AI Paper Atlas

Research map / Large language models / Agents, code and math

Faults in Our Formal Benchmarking: Dataset Defects and Evaluation Failures in Lean Theorem Proving

Pawan Sasanka Ammanamanchi, Siddharth Bhat, Stella Biderman

ICML 2026 · Large language models › Agents, code and math · approach: Retrieval-augmented formal theorem proving · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
manually correct flawed formal statements (20 problems with proven issues)DeepSeek-Prover-V2-7B · 20 corrected benchmark problemsproblems solved: 0/20 -> 3/20Improved
manually correct flawed formal statements (20 problems with proven issues)Kimina-Prover-8B · 20 corrected benchmark problemsproblems solved: 0/20 -> 2/20Improved
repair weakened statements (ProofNet to ProofNet#)provers · ProofNetpass rateHurt
add LLM false-positive filtering to static checkersstatic checkers · ProverBenchnumber of findings: 427 -> 277Improved
use Gemini 3.0 Flash instead of DeepSeek-V3 for warning filteringLLM verifier · ProverBench warning setaccuracy: 68.5% -> 83.3%Improved
LLM semantic audit on specification errors versus formalization errorsSonnet 4.5 + Thinking · 92-problem challenge setF1: 0.51 vs 0.18Mixed
use Sonnet 4.5 + Thinking instead of GPT-5.2 + Thinking for semantic auditsemantic audit · 92-problem challenge setaccuracy: 54.6% -> 68.1%Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy