NeurIPS paper: with mixed data and RL-matched scale, SFT can generalize as well as or better than RL

kastnerkyle · x · 2026-11-17

A NeurIPS 2026 paper, "Rethinking 'RL Generalizes, SFT Memorizes': The Role of SFT Data" (NUS/Yale/Cisco Foundation AI, arXiv:2611.14912), argues the community consensus since DeepSeek-R1 may stem from an unfair comparison:

Original post →

More from Research

Research channel →