arXiv AI By Jianghan Shen, Siqi Luo, Yue Li, Jiyao Liu, Wanying Qu, Yi Zhang, Ziyan Huang, Tianbin Li, Ming Hu, Xiaohong Liu, Yirong Chen, Junjun He

A First-Principles Derivation of LLM Policy Optimization: From Expected Reward to GRPO and Its Structural Extensions

Read the original on arXiv AI →

arXiv:2606. 16733v1 Announce Type: new Abstract: Policy gradient algorithms for language models optimize the same objective $J(\theta) = \mathbb{E}*{\tau \sim p*\theta(\tau)}[R(\tau)]$, which has exactly two factors: the trajectory probability $p_\theta(\tau)$ and the reward $R(\tau)$.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.

arXiv Machine Learning
Jun 5

Selective-Advantage Entropy-Adaptive Horizon GRPO: Asymmetric Token-Level Discounting for Efficient Reinforcement Learning of Language Models

arXiv:2606. 05434v1 Announce Type: new Abstract: Group Relative Policy Optimisation (GRPO) has emerged as an effective reinforcement-learning algorithm for aligning language models on reasoning tasks, but it treats every token position and every sampled rollout symmetrically.

By Chirag Chawla, Rohan Charudatt Salvi, Madhav S. Baidya
Hugging Face Trending Papers
Jul 30

LoRA Scaffolded Policy Optimization (LSPO): A Sampling-Time Low-Rank Scaffold for Recovering Reinforcement-Learning Gradient on Zero-Reward Cliff Prompts

Reinforcement learning from verifiable rewards (RLVR) for mathematical reasoning suffers from a structural blind spot: on "cliff" prompts-those on which every sampled rollout in a group fails-the group-normalized advantage is identically zero, so GRPO produces no gradient on precisely the prompts at the frontier of the model's capability. We introduce LoRA Scaffolded Policy Optimization (LSPO), a sampling-time mechanism that recovers this lost gradient.