uncertainty markers LLM

Uncertainty Markers Double LLM Reasoning Accuracy: The Self-Correction Prompting Technique You Shouldn't Sleep On

Embedding hesitation cues like 'wait, let me verify' into your prompts activates built-in self-correction that chain-of-thought never could