NeurIPS 2022 · 2022
Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia
Google Research
What this paper is
Chain-of-thought prompting (Wei et al., 2022) asks an LLM to write intermediate reasoning steps before its final answer, sharply improving math, logic, and multi-step benchmark accuracy. The gain emerges only at sufficient model scale, and the paper established step-by-step reasoning as a standard capability elicitation technique.
Abstract summary
Demonstrates that prompting LLMs to articulate intermediate reasoning steps before producing a final answer ('chain-of-thought prompting') dramatically improves accuracy on math, logic, and multi-step problem-solving benchmarks. The improvement scales with model size and emerges only at sufficient scale. Establishes step-by-step reasoning as a critical prompting technique and a foundation for later 'reasoning model' designs.
Our summary in our own words, see the canonical source links below for the original abstract.
Why we cite this paper
Chain-of-Thought is the technique that 'reasoning models' (OpenAI o1/o3, Anthropic Claude with extended thinking, DeepSeek R1) train into the model rather than relying on prompting alone. The category emerged directly from this paper's framing. Our /trend/agentic-ai-frameworks-2026 tracks the engineering acceleration in this category.
Where this matters for deal flow
Key findings
- 1Chain-of-thought prompting dramatically improves LLM performance on multi-step reasoning tasks.
- 2The capability emerges only at sufficient model scale (~100B parameters).
- 3Step-by-step reasoning can be elicited via few-shot prompting without model retraining.
- 4Foundation for modern reasoning models that train extended chain-of-thought as a native capability.
Canonical sources
Related glossary terms
Frequently Asked Questions
What is chain-of-thought prompting?▾
A prompting technique where the model is instructed to articulate intermediate reasoning steps before producing a final answer. See /define/chain-of-thought for the full term definition.
Who wrote the chain-of-thought paper?▾
Jason Wei and colleagues at Google Research, published at NeurIPS 2022 (arXiv:2201.11903).
Does chain-of-thought work on small models?▾
No. The paper shows the benefit emerges only at sufficient model scale (around 100B parameters). Below that threshold, step-by-step prompting does not reliably improve reasoning accuracy.
How does this relate to reasoning models?▾
Modern reasoning models (OpenAI o1/o3, Claude with extended thinking, DeepSeek R1) train extended chain-of-thought into the model as a native capability, rather than relying on few-shot prompting alone.
See who is building on this, before the round prices it in
We track engineering acceleration across the AI & Machine Learning sector this paper informs: commit velocity, contributor influx, and repo-creation pulse, surfacing breakout teams 21 to 47 days before the fundraise is public.
Five breakout startups, every Sunday, before the round gets crowded
The free Acceleration Watch: five venture-backed teams accelerating on the engineering signal, translated into plain English, 21 to 47 days before the deck circulates. No code-reading, no card.
Other research papers
NeurIPS 2017 · 2017
Attention Is All You Need
NeurIPS 2020 · 2020
Language Models are Few-Shot Learners
NeurIPS 2022 · 2022
Training language models to follow instructions with human feedback
NeurIPS 2020 · 2020
Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
ICLR 2022 · 2021
LoRA: Low-Rank Adaptation of Large Language Models
arXiv preprint · 2022
Constitutional AI: Harmlessness from AI Feedback
Read our own methodology paper
Code-Side Sourcing methodology, replicable on the open dataset.
Read /methodologyRelated papers
More in Research Paper