AI Paper Atlas

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

Rewriting Pre-Training Data Boosts LLM Performance in Math and Code

Kazuki Fujii, Yukito Tajima, Sakae Mizuki, Masaki Kawamura, Hinari Shimada, Taihei Shiotani, Koshiro Saito, Masanari Oi, Taishi Nakamura, Takumi Okamoto, Shigeki Ishida, Kakeru Hattori and 5 more

ICLR 2026 · Large language models › Agents, code and math · approach: Execution-based code verification · 30 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
add LLM style-guided rewriting (SGCR) after syntax and lint filtering (ablation)SwallowCode pipeline · HumanEval/HumanEval+pass@1: over 7-9 pointsImproved
add self-contained optimization rewriting (SCOR) after SGCR (ablation)SwallowCode pipeline · HumanEval/HumanEval+pass@1: over 5-6 pointsImproved
add syntax error filtering (ablation)The-Stack-v2 Python · HumanEval/HumanEval+pass@1Improved
add pylint filtering (threshold 7.0) (ablation)The-Stack-v2 Python · HumanEval/HumanEval+pass@1: more than 1 pointImproved
add LLM-based score filtering over linter filtering (ablation)The-Stack-v2 Python · HumanEval/HumanEval+pass@1: less than 1 pointNo clear effect
apply SGCR after filtering instead of on raw data (ablation)SGCR · HumanEval/HumanEval+pass@1: +0.4-2.1 pointsImproved
apply SGCR (renaming functions) (ablation)SwallowCode · MBPPaccuracy: approximately 10-point decreaseHurt
replace Stack-Edu with SwallowCodeLlama-3.1-8B · HumanEvalpass@1: +17.0Improved
replace Stack-Edu with SwallowCodeQwen2-7B · HumanEvalpass@1: 39.3 -> 49.6Improved
replace Finemath-4+ with SwallowMathLlama-3.1-8B · GSM8Kaccuracy: +12.4Improved
replace Finemath-4+ with SwallowMathLlama-3.1-8B · MATHaccuracy: +7.6Improved
switch from Japanese-heavy to English-heavy mixtureLlama-3.1-8B · HumanEvalaccuracy: 0.583 (English-heavy mixture)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