Beyond Training Objectives: Interpreting Reward Model Divergence in Large Language Models

Large language models (LLMs) fine-tuned by reinforcement learning from human feedback (RLHF) are becoming more widely deployed. We coin the term $\textit{Implicit Reward Model}$ (IRM) to refer to the changes that occur to an LLM during RLHF that result in high-reward generations. We interpret IRMs, and measure their divergence from the RLHF reward model used in the fine-tuning process that induced them. By fitting a linear function to an LLM's IRM, a reward model with the same type signature as …

beyond cs.lg divergence feedback human human feedback language language models large language large language models llm llms reinforcement reinforcement learning reward model rlhf training