Cannot-Self-Correct tests the strong claim that LLMs can revise their own reasoning answers without any external signal ...

Cannot-Self-Correct tests the strong claim that LLMs can revise their own reasoning answers without any external signal about correctness. Across three benchmarks (GSM8K, CommonSenseQA, HotPotQA), the answer is no: the model's confidence carries over from the initial answer into the revision, and the self-correction loop tends to degrade rather than improve performance. The result refutes the class of approach Self-Refine belongs to.https://benjaminhan.net/posts/20260516-cannot-self-correct/?utm_source=mastodon&utm_medium=social#LLMs #AI #Reasoning #Metacognition

Read Original

Related