View of AutoCrit - A Meta-Reasoning Framework for Self-Critique and Iterative Error Correction in LLM Chains-of-Thought: Attributing Performance to Model Components

Return to Issue Details AutoCrit - A Meta-Reasoning Framework for Self-Critique and Iterative Error Correction in LLM Chains-of-Thought: Attributing Performance to Model Components Download Download PDF