PILAF: Optimal Human Preference Sampling for Reward Modeling
This addresses the challenge of improving human preference sampling for reward modeling in RLHF, which is crucial for real-world AI alignment, though it appears incremental as it builds on existing RLHF frameworks.
The paper tackles the problem of aligning large language models with human values by proposing PILAF, a novel response sampling strategy for preference labeling that optimizes reward modeling in RLHF, achieving strong performance in iterative and online settings.
As large language models increasingly drive real-world applications, aligning them with human values becomes paramount. Reinforcement Learning from Human Feedback (RLHF) has emerged as a key technique, translating preference data into reward models when oracle human values remain inaccessible. In practice, RLHF mostly relies on approximate reward models, which may not consistently guide the policy toward maximizing the underlying human values. We propose Policy-Interpolated Learning for Aligned Feedback (PILAF), a novel response sampling strategy for preference labeling that explicitly aligns preference learning with maximizing the underlying oracle reward. PILAF is theoretically grounded, demonstrating optimality from both an optimization and a statistical perspective. The method is straightforward to implement and demonstrates strong performance in iterative and online RLHF settings where feedback curation is critical.