arXiv:2607. 17524v1 Announce Type: cross Abstract: We propose Token-Level Off-Policy Labeling (TOPL), an off-policy training paradigm that reframes post-training as a token-level correctness prediction task.
By Zitong Huang, Gustavo Lucas Carvalho, Deqing Fu, Robin Jia
We propose Token-Level Off-Policy Labeling (TOPL), an off-policy training paradigm that reframes post-training as a token-level correctness prediction task. Our key intuition is that by training the model to distinguish good and bad tokens in a response, we naturally guide the model towards generating good tokens, while avoiding the pitfalls that come with directly training the model to generate off-policy tokens.
The paper introduces PLUS, a framework that uses reinforcement learning to generate text-based summaries of individual users’ preferences, characteristics, and past conversations. These summaries condition a reward model, allowing it to predict personalized response preferences and improving reward accuracy by 11–77 % over the standard Bradley‑Terry model. PLUS demonstrates robust performance with new users and topics, achieves a 25 % improvement over existing personalized RLHF techniques, and enables zero‑shot personalization for state‑of‑the‑art models like GPT‑4.
By Hyunji Nam, Yanming Wan, Mickel Liu, Peter Ahnn, Jianxun Lian, Natasha Jaques
arXiv:2608. 09605v1 Announce Type: cross Abstract: Large Language Models (LLMs) have emerged as powerful tools for improving recommendation systems.
By Wenqiao Zhu, Chao Xu, Haipang Wu, Ji Liu
arXiv:2602. 13940v2 Announce Type: replace-cross Abstract: Tokenization is a hardcoded compression step which remains in the training pipeline of Large Language Models (LLMs), despite a general trend towards architectures becoming increasingly end-to-end.
By Sam Dauncey, Roger Wattenhofer
arXiv:2607. 04728v1 Announce Type: cross Abstract: Reinforcement learning (RL) post-training for large language models (LLMs) follows a efficient paradigm of "rollout then update", which inevitably results in off-policy training data.
By Yu Li, Xiuyu Li, Mingyang Yi, Jiaxing Wang, zhangliangxu, Zhaolong Xing, Zhen Chen