4 papers
BiPACE: Bisimulation-Guided Policy Optimization with Action Counterfactual Estimation for LLM Agents
Hanyang Wang, Weijieying Ren, Yuxiang Zhang +4
Stepwise group-based RL is an attractive way to train long-horizon LLM agents without a learned critic: it reuses multiple sampled rollouts to estimate local advantages. Its weakne…
Learning to Foresee: Unveiling the Unlocking Efficiency of On-Policy Distillation
Yuchen Cai, Ding Cao, Liang Lin +9
On-policy distillation (OPD) has emerged as an efficient post-training paradigm for large language models. However, existing studies largely attribute this advantage to denser and…
On Predictability of Reinforcement Learning Dynamics for Large Language Models
Yuchen Cai, Ding Cao, Xin Xu +7
Recent advances in reasoning capabilities of large language models (LLMs) are largely driven by reinforcement learning (RL), yet the underlying parameter dynamics during RL trainin…
On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models
Ding Cao, Yuchen Cai, Yuqing Huang +4
Sequential knowledge editing techniques aim to continuously update knowledge in large language models at low cost, preventing models from generating outdated or incorrect informati…