self-consistency-multiple-cot-majority-vote
IN premise — entries/2026/06/21/wiki-Prompt_engineering-chunk-2.md
Created 2026-06-21T09:50:10+00:00
Self-consistency improves Chain-of-Thought by sampling multiple reasoning paths and selecting the most common conclusion via majority vote (Wang et al., ICLR 2023, arXiv:2203.11171)
Summary
Rather than trusting a single line of reasoning from a language model, you can ask it to work through a problem several different ways and then adopt whichever final answer shows up most often. This matters because it turns one-shot reasoning into a more robust, self-correcting process without needing a separate judge or reward model, giving the system a cheap, empirically validated way to boost answer reliability.
Dependents
These beliefs depend on this one:
- OUT inference-compute-reliably-improves-reasoning — Self-consistency (sampling multiple chain-of-thought paths and selecting by majority vote) and Tree of Thoughts (branching reasoning with backtracking) independently demonstrate that LLM reasoning reliability improves with inference-time compute investment — establishing a general compute-for-quality tradeoff at inference time — unless prompt sensitivity means each individual reasoning path is itself unreliable enough to undermine the ensemble.
- IN structured-reasoning-prompting-evolved-from-linear-to-branching — Prompting for reasoning evolved from linear chain-of-thought (single path) to self-consistency (multiple paths, majority vote) to tree-of-thoughts (branching with backtracking), progressively adding search structure.