Chain-of-Thought Prompting vs Self-Consistency
Chain-of-thought prompting asks a model to write intermediate reasoning steps before its answer, in a single response. Self-consistency builds on it by sampling several chain-of-thought responses to the same question and returning the most common final answer, which improves accuracy at the cost of several times the inference.
Definitions
- Chain-of-Thought Prompting PTL-0024
- Chain-of-thought (CoT) prompting elicits a sequence of intermediate reasoning steps from a language model before its final answer, which improves performance on multi-step arithmetic, commonsense, and symbolic reasoning tasks.
- Self-Consistency PTL-0028
- Self-consistency samples multiple chain-of-thought reasoning paths for the same question and returns the answer that appears most often, rather than relying on a single greedy decode.
Differences
| Chain-of-Thought Prompting | Self-Consistency | |
|---|---|---|
| What it changes | The prompt or instruction | How many responses are sampled and how the answer is chosen |
| Model calls per question | One | Several (one per sampled reasoning path) |
| Final answer | The answer at the end of the single chain | The majority answer across chains |
| Cost | Baseline plus reasoning tokens | Roughly the baseline multiplied by the number of samples |
Which to use
- Use chain-of-thought alone when one careful response is enough or cost and latency matter.
- Add self-consistency when answers are short and comparable, such as a number or a choice, and accuracy is worth extra calls.
Sources
- Wei et al. (2022). Chain-of-Thought Prompting Elicits Reasoning in Large Language Models.
- Wang et al. (2022). Self-Consistency Improves Chain of Thought Reasoning in Language Models.
Cite this page
Protologue. (2026). Chain-of-Thought Prompting vs Self-Consistency. In Protologue: A Taxonomy of Prompting and LLM Techniques (v1.4.0). https://protologue.com/compare/chain-of-thought-vs-self-consistency/