Distillation and Refinement of Reasoning in Small Language Models for Document Re-ranking
This provides a scalable and interpretable solution for information retrieval systems, addressing the need for efficient reasoning in small language models without expensive human annotations.
The paper tackles the problem of training small language models for document re-ranking by combining knowledge distillation and reinforcement learning to generate high-quality training examples automatically, resulting in a 3B parameter model that achieves state-of-the-art performance on the BRIGHT benchmark, ranking third while using fewer parameters than larger models.
We present a novel approach for training small language models for reasoning-intensive document ranking that combines knowledge distillation with reinforcement learning optimization. While existing methods often rely on expensive human annotations or large black-box language models, our methodology leverages web data and a teacher LLM to automatically generate high-quality training examples with relevance explanations. By framing document ranking as a reinforcement learning problem and incentivizing explicit reasoning capabilities, we train a compact 3B parameter language model that achieves state-of-the-art performance on the BRIGHT benchmark. Our model ranks third on the leaderboard while using substantially fewer parameters than other approaches, outperforming models that are over 20 times larger. Through extensive experiments, we demonstrate that generating explanations during inference, rather than directly predicting relevance scores, enables more effective reasoning with smaller language models. The self-supervised nature of our method offers a scalable and interpretable solution for modern information retrieval systems.