作者位次 #2/2ACL 2026Findings· Findings of ACL
CRPS : Curriculum Replay via Progressive Suffixes from Successful Trajectories for Long-Horizon LLM Agents
Zijing Zhang · Xiajie Yang
摘要
Long-horizon LLM agents trained with sparse terminal rewards tend to experience slow and unstable learning, and the issue is amplified by group-normalized on-policy objectives commonly used for LLM training (e.g., GRPO). When rollout groups collapse to nearly all failures early in training, withi...