Beyond Output Critique: Self-Correction via Task Distillation

2026-01-31

AI Summary

This work proposes SELF-THOUGHT, a novel self-correction framework that addresses the limitations of existing large language model (LLM) correction methods, which often focus only on superficial errors and fail to rectify deep reasoning flaws. By distilling complex reasoning tasks into structured templates through task distillation, SELF-THOUGHT guides the model to reconstruct its reasoning process before generating a final answer, enabling high-quality self-correction without fine-tuning or external verifiers. The framework supports cross-model template transfer, allowing smaller models to leverage abstract knowledge distilled from larger ones. Experimental results demonstrate that SELF-THOUGHT significantly enhances accuracy, robustness, and generalization across diverse reasoning tasks for both large and small models, offering a scalable new paradigm for self-correcting language systems.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Computer Vision: Large Vision Models

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large language models (LLMs) have shown promising self-correction abilities, where iterative refinement improves the quality of generated responses. However, most existing approaches operate at the level of output critique, patching surface errors while often failing to correct deeper reasoning flaws. We propose SELF-THOUGHT, a framework that introduces an intermediate step of task abstraction before solution refinement. Given an input and an initial response, the model first distills the task into a structured template that captures key variables, constraints, and problem structure. This abstraction then guides solution instantiation, grounding subsequent responses in a clearer understanding of the task and reducing error propagation. Crucially, we show that these abstractions can be transferred across models: templates generated by larger models can serve as structured guides for smaller LLMs, which typically struggle with intrinsic self-correction. By reusing distilled task structures, smaller models achieve more reliable refinements without heavy fine-tuning or reliance on external verifiers. Experiments across diverse reasoning tasks demonstrate that SELF-THOUGHT improves accuracy, robustness, and generalization for both large and small models, offering a scalable path toward more reliable self-correcting language systems.

Problem

• self-correction
• reasoning flaws
• output critique
• task abstraction
• error propagation

Innovation

• task distillation
• self-correction
• task abstraction
• structured reasoning
• cross-model transfer