PLOT: Enhancing Preference Learning via Optimal Transport
This work addresses the challenge of enhancing preference learning in LLMs for better alignment with human values and logic, representing an incremental improvement over existing methods.
The paper tackles the problem of limited performance gains and high computational costs in preference learning for Large Language Models by introducing PLOT, a method that uses optimal transport for token-level loss, which improved alignment performance across seven subpreferences while maintaining fluency and coherence.
Preference learning in Large Language Models (LLMs) has advanced significantly, yet existing methods remain limited by modest performance gains, high computational costs, hyperparameter sensitivity, and insufficient modeling of global token-level relationships. We introduce PLOT, which enhances Preference Learning in fine-tuning-based alignment through a token-level loss derived from Optimal Transport. By formulating preference learning as an Optimal Transport Problem, PLOT aligns model outputs with human preferences while preserving the original distribution of LLMs, ensuring stability and robustness. Furthermore, PLOT leverages token embeddings to capture semantic relationships, enabling globally informed optimization. Experiments across two preference categories - Human Values and Logic & Problem Solving - spanning seven subpreferences demonstrate that PLOT consistently improves alignment performance while maintaining fluency and coherence. These results substantiate optimal transport as a principled methodology for preference learning, establishing a theoretically grounded framework that provides new insights for preference learning of LLMs.