AutoPRM: Automating Procedural Supervision for Multi-Step Reasoning via Controllable Question Decomposition
This addresses the bottleneck of manual feedback in reasoning tasks for AI researchers and practitioners, though it is an incremental advancement building on existing reinforcement learning and decomposition techniques.
The paper tackles the problem of multi-step reasoning in large language models by automating procedural supervision, eliminating the need for extensive manual labeling. The result is a significant performance improvement on mathematical and commonsense reasoning tasks over state-of-the-art methods.
Recent advancements in large language models (LLMs) have shown promise in multi-step reasoning tasks, yet their reliance on extensive manual labeling to provide procedural feedback remains a significant impediment. To address this challenge, in this paper, we propose a novel self-supervised framework AutoPRM that efficiently enhances the fine-tuning of LLMs for intricate reasoning challenges. Specifically, AutoPRM first decomposes complex problems into more manageable subquestions with a controllable granularity switch, then sequentially apply reinforcement learning to iteratively improve the subquestion solver. Additionally, we propose context-guided-decoding to avoid reward tampering and guide the subquestion solver towards the solution of the holistic problem. Extensive experiments show that AutoPRM significantly improves performance on mathematical and commonsense reasoning tasks over SOTA. More encouragingly, AutoPRM can be easily integrated with other orthogonal reasoning pipelines.