Prompt Engineering · Established · Intermediate

Self-Consistency Prompting

Also known as: Self-Consistency Decoding, CoT Majority Vote

A decoding technique that samples multiple diverse Chain of Thought reasoning paths from an LLM and selects the most consistent answer.

What Self-Consistency Prompting is

Self-Consistency significantly improves accuracy on math, logic, and multi-step reasoning prompts over single greedy decoding.

How it works

Prompts the model with temperature > 0 to generate N candidate reasoning chains, then aggregates final answers by frequency.

Why it matters

Used in production automated evaluation pipelines and benchmark verification workflows.

Common uses