SpecRoll: Fast-Slow Verifier-Feedback Adaptation for Speculative Reinforcement Learning Rollouts
This work introduces SpecRoll, a speculative rollout engine that preserves the target model's sampling distribution while adapting at two timescales, and combines concurrency-aware sparse-tree verification and exact target verification, leaving the target rollout distribution and GRPO objective unchanged.
Nhật Minh Phạm, Duy-Tung Doan, Thi Duyen Ngo et al.
· 1 citation
· ⚡1