IPO: Your Language Model is Secretly a Preference Classifier

TL;DR

This paper introduces Implicit Preference Optimization (IPO), a novel method that uses large language models as preference classifiers to reduce reliance on external feedback, achieving comparable performance to traditional RLHF methods.

Contribution

The paper proposes IPO, a new approach that leverages LLMs as preference classifiers, reducing costs and dependence on external reward models in aligning language models.

Findings

Models trained with IPO perform comparably to reward model-based methods.

LLMs can effectively serve as preference classifiers for self-improvement.

Evaluation on RewardBench confirms IPO's effectiveness across model sizes and architectures.

Abstract

Reinforcement learning from human feedback (RLHF) has emerged as the primary method for aligning large language models (LLMs) with human preferences. While it enables LLMs to achieve human-level alignment, it often incurs significant computational and financial costs due to its reliance on training external reward models or human-labeled preferences. In this work, we propose Implicit Preference Optimization (IPO), an alternative approach that leverages generative LLMs as preference classifiers, thereby reducing the dependence on external human feedback or reward models to obtain preferences. We conduct a comprehensive evaluation on the preference classification ability of LLMs using RewardBench, assessing models across different sizes, architectures, and training levels to validate our hypothesis. Furthermore, we investigate the self-improvement capabilities of LLMs by generating multiple responses for a given instruction and employing the model itself as a preference classifier for Direct Preference Optimization (DPO)-based training. Our findings demonstrate that models trained through IPO achieve performance comparable to those utilizing state-of-the-art reward models for obtaining preferences.