paper-with-me

Papers

Greedy Sampling Is Provably Efficient for RLHF

2025-10-28 · Di Wu, Chengshuai Shi, Jing Yang, Cong Shen arxiv

Reinforcement Learning from Human Feedback (RLHF) has emerged as a key technique for post-training large language models. Despite its empirical success, the theoretical understanding of RLHF is still limited, as learning the KL-regularized target with only preference feedback poses additional challenges compared with canonical RL. Existing works mostly study the reward-based Bradley-Terry (BT) preference model, and extend classical designs utilizing optimism or pessimism. This work, instead, considers the general preference model (whose practical relevance has been observed recently) and obtains performance guarantees with major, order-wise improvements over existing ones. Surprisingly, these results are derived from algorithms that directly use the empirical estimates (i.e., greedy sampling), as opposed to constructing optimistic or pessimistic estimates in previous works. This insight has a deep root in the unique structural property of the optimal policy class under the KL-regularized target, and we further specialize it to the BT model, highlighting the surprising sufficiency of greedy sampling in RLHF.

📄 PDF Abstract BibTeX arXiv:2510.24700

Code (0)

등록된 구현이 없습니다.

Tasks

Reinforcement Learning

Similar Papers 제목 키워드 기반

Provably Efficient Regularized Online RLHF with Generalized Bilinear Preferences

2026-02-26 · Junghyun Lee, Minju Hong, Kwang-Sung Jun, Chulhee Yun 외 arxiv

We consider the problem of regularized best-response max-regret minimization in online RLHF under general preferences and bandit feedback. While various regularizers are utilized to robustify alignment, known polylogarit…

SharedRep-RLHF: A Shared Representation Approach to RLHF with Diverse Preferences

2025-09-03 · Arpan Mukherjee, Marcello Bullo, Deniz Gündüz arxiv

Uniform-reward reinforcement learning from human feedback (RLHF), which trains a single reward model to represent the preferences of all annotators, fails to capture the diversity of opinions across sub-populations, inad…

Reinforcement Learning

Provably Efficient RLHF Pipeline: A Unified View from Contextual Bandits

2025-02-11 · Long-Fei Li, Yu-Yang Qian, Peng Zhao, Zhi-Hua Zhou

Reinforcement Learning from Human Feedback (RLHF) is a widely used approach for aligning Large Language Models (LLMs) with human preferences. While recent advancements have provided valuable insights into various stages …

Computational EfficiencyMulti-Armed Bandits

Hybrid Preference Optimization for Alignment: Provably Faster Convergence Rates by Combining Offline Preferences with Online Exploration

2024-12-13 · Avinandan Bose, Zhihan Xiong, Aadirupa Saha, Simon Shaolei Du 외

Reinforcement Learning from Human Feedback (RLHF) is currently the leading approach for aligning large language models with human preferences. Typically, these models rely on extensive offline preference datasets for tra…

Accelerating RLHF Training with Reward Variance Increase

2025-05-29 · Zonglin Yang, Zhexuan Gu, HouDuo Qi, Yancheng Yuan

Reinforcement learning from human feedback (RLHF) is an essential technique for ensuring that large language models (LLMs) are aligned with human values and preferences during the post-training phase. As an effective RLH…