Search papers, labs, and topics across Lattice.
This paper addresses reward overoptimization in Reinforcement Learning from Human Feedback (RLHF) by introducing Real-Time Aligned Reward Model (R2M). R2M enhances reward models by incorporating real-time feedback from the evolving hidden states of the policy model, going beyond reliance on surface semantic information. The approach mitigates reward discrepancy caused by policy distribution shifts during RL, leading to improved alignment between the reward model and policy model.
Stop overfitting your reward model: R2M leverages real-time policy feedback to dynamically align the reward model with the evolving policy distribution, reducing reward overoptimization in RLHF.
Reinforcement Learning from Human Feedback (RLHF) is a pivotal technique for aligning large language models (LLMs) with human preferences, yet it is susceptible to reward overoptimization, in which policy models overfit to the reward model, exploit spurious reward patterns instead of faithfully capturing human intent. Prior mitigations primarily relies on surface semantic information and fails to efficiently address the misalignment between the reward model (RM) and the policy model caused by continuous policy distribution shifts. This inevitably leads to an increasing reward discrepancy, exacerbating reward overoptimization. To address these limitations, we introduce R2M (Real-Time Aligned Reward Model), a novel lightweight RLHF framework. R2M goes beyond vanilla reward models that solely depend on the semantic representations of a pretrained LLM. Instead, it leverages the evolving hidden states of the policy (namely policy feedback) to align with the real-time distribution shift of the policy during the RL process. This work points to a promising new direction for improving the performance of reward models through real-time utilization of feedback from policy models.
