Shobhita Sundaram(@shobsund):Can a model learn to break its own reasoning plateau? In our new paper, we show that LLMs can be taught with meta-RL to generate their own "stepping stones" that kickstart learning on hard math problems (0/128 success rate) where direct RL fails. Paper 📝: https://t.co/lUlrJt6bwq Blog post 🌐: https://t.co/v1y24h1fP4 (1/n)

Shobhita Sundaram

@shobsund

PhD student at @MIT_CSAIL working on self-improvement and synthetic data || prev at @GoogleDeepMind @AIatMeta

가입 August 2022

435 팔로잉 중 1.8K 팬

Shobhita Sundaram@shobsund

2026.01.27 15:35

Can a model learn to break its own reasoning plateau? In our new paper, we show that LLMs can be taught with meta-RL to generate their own "stepping stones" that kickstart learning on hard math problems (0/128 success rate) where direct RL fails. Paper 📝: Blog post 🌐: (1/n)