reasoning-correction

The paper introduces DCGC, a Masked Diffusion Model framework for globally correcting flawed reasoning traces in LLMs by conditioning on an imperfect draft. It improves accuracy on reasoning benchmarks without ground-truth failure labels.

Deep Interaction proposes a method for directly editing erroneous reasoning steps in chain-of-thought outputs from large language models, achieving over 25% improvement in correction success and 40% reduction in token usage.