all AI news
Mitigating Reward Hacking via Information-Theoretic Reward Modeling
Feb. 15, 2024, 5:42 a.m. | Yuchun Miao, Sen Zhang, Liang Ding, Rong Bao, Lefei Zhang, Dacheng Tao
cs.LG updates on arXiv.org arxiv.org
Abstract: Despite the success of reinforcement learning from human feedback (RLHF) in aligning language models with human values, reward hacking, also termed reward overoptimization, remains a critical challenge, which primarily stems from limitations in reward modeling, i.e., generalizability of the reward model and inconsistency in the preference dataset. In this work, we tackle this problem from an information theoretic-perspective, and propose a generalizable and robust framework for reward modeling, namely InfoRM, by introducing a variational information …
abstract arxiv challenge cs.ai cs.lg dataset feedback hacking human human feedback information language language models limitations modeling reinforcement reinforcement learning reward model rlhf success type values via
More from arxiv.org / cs.LG updates on arXiv.org
Jobs in AI, ML, Big Data
Founding AI Engineer, Agents
@ Occam AI | New York
AI Engineer Intern, Agents
@ Occam AI | US
AI Research Scientist
@ Vara | Berlin, Germany and Remote
Data Architect
@ University of Texas at Austin | Austin, TX
Data ETL Engineer
@ University of Texas at Austin | Austin, TX
Machine Learning Research Scientist
@ d-Matrix | San Diego, Ca