Loading timeline…
20222020s
Chain-of-Thought Prompting
Forcing standard language models to elicit step-by-step logical reasoning before answering.
Why It Was Important
Jason Wei and colleagues observed that simply appending the phrase 'Let's think step-by-step' drastically increased an LLM's ability to solve math and logic problems. By generating intermediate reasoning tokens in the context window, the model avoided single-pass estimation, unlocking latent problem-solving capabilities.
Who Invented It
Jason Wei et al. (Google Brain)
NLP researchers focusing on emergent behaviors of scaling.
Applications
- AI Prompt Engineering
- Mathematical Problem Solving
- Autonomous Agents
Key Papers
- Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
Jason Wei et al. · NeurIPS 2022