method · first seen 4 Oct 2026
clipped policy ratio
Also: PPO clipping, proximal policy optimization clipping for language models
Limiting the probability ratio between new and old policies during policy gradient updates to stabilize training.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Prompt Dominance and Asymmetric Verifier Costs: Empirical Ablations of GRPO at 1B Scale on GSM8KPapers · 4 Oct 2026 · Yi Hou
- EPIC: Rethinking Trust Regions for Autoregressive LLM Policies via Exact-KL Prefix Importance-Ratio ClippingPapers · 21 Oct 2026 · Chengxi Li, Hanlin Wu, Zhicheng Cai, Huan-ang Gao, Wei-Ying Ma, Jingjing Liu, Ha…
Latest sources
- EPIC: Rethinking Trust Regions for Autoregressive LLM Policies via Exact-KL Prefix Importance-Ratio ClippingPapers · 21 Oct 2026 · Chengxi Li, Hanlin Wu, Zhicheng Cai, Huan-ang Gao, Wei-Ying Ma, Jingjing Liu, Ha…
- Prompt Dominance and Asymmetric Verifier Costs: Empirical Ablations of GRPO at 1B Scale on GSM8KPapers · 4 Oct 2026 · Yi Hou