Structured Reasoning for Large Language Models
This addresses inefficiencies in reasoning for users of large language models, though it is incremental as it builds on existing paradigms like Generate-Verify-Revise.
The paper tackles the problem of redundant and ineffective reasoning steps in large language models by proposing Structured Reasoning (SCR), a framework that decouples reasoning trajectories into trainable components, resulting in up to a 50% reduction in output token length.
Large language models (LLMs) achieve strong performance by generating long chains of thought, but longer traces always introduce redundant or ineffective reasoning steps. One typical behavior is that they often perform unnecessary verification and revisions even if they have reached the correct answers. This limitation stems from the unstructured nature of reasoning trajectories and the lack of targeted supervision for critical reasoning abilities. To address this, we propose Structured Reasoning (SCR), a framework that decouples reasoning trajectories into explicit, evaluable, and trainable components. We mainly implement SCR using a Generate-Verify-Revise paradigm. Specifically, we construct structured training data and apply Dynamic Termination Supervision to guide the model in deciding when to terminate reasoning. To avoid interference between learning signals for different reasoning abilities, we adopt a progressive two-stage reinforcement learning strategy: the first stage targets initial generation and self-verification, and the second stage focuses on revision. Extensive experiments on three backbone models show that SCR substantially improves reasoning efficiency and self-verification. Besides, compared with existing reasoning paradigms, it reduces output token length by up to 50%.