COLM 2026 · Equal contribution
Publications
Research on training data, foundation models, and reasoning. Google Scholar ↗
2026
2025
COLM 2025 · Equal contribution
MegaMath: Pushing the Limits of Open Math Corpora
ICML 2025 · Equal contribution
Programming Every Example: Lifting Pre-training Data Quality like Experts at Scale
AI4Math @ ICML 2025 · Equal contribution
OctoThinker: Mid-training Incentivizes Reinforcement Learning Scaling
2024
NeurIPS Datasets & Benchmarks 2024
MathPile: A Billion-Token-Scale Pretraining Corpus for Math
NeurIPS Datasets & Benchmarks 2024
OlympicArena: Benchmarking Multi-discipline Cognitive Reasoning for Superintelligent AI
COLM 2024
Is ChatGPT a Good Sentiment Analyzer?
ACL 2024 · Equal contribution
Ask Again, Then Fail: Large Language Models' Vacillations in Judgment
Preprint · Equal contribution