Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Teach Yourself Where to Look: On-Policy Attention Self-Distillation for Reasoning
Safaeid Hossain Arib, Rabeya Akter, Ismam Nur Swapnil +3
On-policy self-distillation trains reasoning models on their own trajectories using dense token distribution guidance from a privileged teacher with access to a verified solution.…
cs.LG2026
Skip What You Can Predict: Predictive Repositioning for Policy Optimization for Efficient LLM Training
Ismam Nur Swapnil, Aranya Saha, Tanvir Ahmed Khan +3
Reinforcement learning with verifiable rewards (RLVR) can improve the reasoning ability of large language models, but repeatedly updating a policy on the same rollout batch is expe…