GRPO-RM: Fine-Tuning Representation Models via GRPO-Driven Reinforcement Learning

TL;DR

This paper introduces GRPO-RM, a reinforcement learning approach for fine-tuning representation models, adapting the GRPO method from language models to improve their performance on real-world datasets.

Contribution

We extend the GRPO reinforcement learning technique to representation models, proposing a new output grouping and reward function for effective post-training optimization.

Findings

GRPO-RM improves representation model performance on multiple datasets.

The method effectively replaces token sampling with output grouping.

Experimental results validate the approach's effectiveness.

Abstract

The Group Relative Policy Optimization (GRPO), a reinforcement learning method used to fine-tune large language models (LLMs), has proved its effectiveness in practical applications such as DeepSeek-R1. It raises a question whether GRPO can be generalized to representation learning models. In this paper, we propose Group Relative Policy Optimization for Representation Model (GRPO-RM), and investigate the performance of GRPO-like policy in post-training representation models. Specifically, our method establishes a predefined output set to functionally replace token sequence sampling in LLMs, thereby generating an output group, which is essential for the probability-driven optimization of GRPO. In addition, a specialized reward function is designed to accommodate the properties of representation models. Extensive experiments are conducted on various real-world datasets to validate the effectiveness of our proposed method.