DiffCoT: Diffusion-styled Chain-of-Thought Reasoning in LLMs
This addresses robustness issues in multi-step mathematical problem-solving for LLMs, though it is an incremental improvement over existing CoT methods.
The paper tackled the problem of error accumulation in Chain-of-Thought reasoning for large language models by proposing DiffCoT, a diffusion-styled framework that reformulates reasoning as an iterative denoising process, resulting in consistent performance improvements on three multi-step reasoning benchmarks.
Chain-of-Thought (CoT) reasoning improves multi-step mathematical problem solving in large language models but remains vulnerable to exposure bias and error accumulation, as early mistakes propagate irreversibly through autoregressive decoding. In this work, we propose DiffCoT, a diffusion-styled CoT framework that reformulates CoT reasoning as an iterative denoising process. DiffCoT integrates diffusion principles at the reasoning-step level via a sliding-window mechanism, enabling unified generation and retrospective correction of intermediate steps while preserving token-level autoregression. To maintain causal consistency, we further introduce a causal diffusion noise schedule that respects the temporal structure of reasoning chains. Extensive experiments on three multi-step CoT reasoning benchmarks across diverse model backbones demonstrate that DiffCoT consistently outperforms existing CoT preference optimization methods, yielding improved robustness and error-correction capability in CoT reasoning.