A Zeroth-Order Paradigm for LLM Preference Alignment
This research provides a new method for aligning large language models with human preferences, which is a key challenge in AI development.
This paper introduces ComPO, a new method for aligning large language models with human preferences.
ComPO is a zeroth-order alignment method that extracts directional information from preference pairs without directly optimizing a differentiable loss.
Experiments on Mistral, Llama, Gemma-2, Qwen3, and Gemma-3 models show improvements over existing direct alignment methods.
Direct preference alignment methods are widely used to align large language models (LLMs) with human preferences because of their computational and memory efficiency. However, likelihood displacement motivates alternative ways to extract information from preference pairs with small likelihood margins. In this paper, we propose and analyze Comparison-based Preference Optimization (ComPO), a zeroth-order alignment method based on comparison oracles. ComPO extracts directional information from these pairs without directly optimizing a differentiable preference loss on them. We establish a convergence guarantee for its basic offline scheme under smoothness, gradient sparsity, and compatibility between the oracle and a latent objective. We further introduce online ComPO, which retains the offline comparison mechanism and uses unlabeled policy generations for reverse-KL control relative to a reference policy. Following the coverage perspective of preference fine-tuning, we establish a performance guarantee for a basic constrained scheme under local coverage and in-distribution pairwise reward accuracy. Experiments on Mistral, Llama, Gemma-2, Qwen3, and Gemma-3 models demonstrate improvements over existing direct alignment methods, including length-controlled win rates, with pair-level diagnostics providing evidence consistent with mitigating likelihood displacement.
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include ComPO, preference alignment, zeroth-order optimization.
Source attribution: curated from arXiv cs.AI (@Peter Chen, Xi Chen, Wotao Yin, Tianyi Lin).
