Robust Multi-Objective Preference Alignment with Online DPO
This work addresses the problem of multi-objective preference alignment for LLMs, offering a more efficient and steerable solution, though it appears incremental as it builds on existing DPO methods.
The paper tackles the challenge of aligning large language models with multiple, potentially conflicting human preferences to enable configurable and personalized AI systems, introducing the MO-ODPO algorithm that Pareto-dominates existing baselines on two benchmarks.
Multi-objective preference alignment of large language models (LLMs) is critical for developing AI systems that are more configurable, personalizable, helpful, and safe. However, optimizing model outputs to satisfy diverse objectives with variable weights at inference time for truly personalized models presents a significant challenge. Existing approaches are either computationally expensive to train or do not sufficiently steer model behaviors. This paper introduces the Multi-Objective Online DPO (MO-ODPO) algorithm, designed to robustly and efficiently align model behaviors with multiple, potentially conflicting human preferences. Our approach incorporates a prompt conditioning mechanism, allowing us to train a single preference-conditional policy, that can adapt to new preference combinations at inference. Experiments on two popular benchmarks show that MO-ODPO Pareto-dominates existing baselines while providing excellent inference-time steerability between diverse objectives.