#72 · Primary category: Prompt Engineering Tools

auto-cot

chain-of-thought gpt-3 gpt3-prompts gpt3-resources large-language-models prompt-engineering reasoning

Official implementation for "Automatic Chain of Thought Prompting in Large Language Models" (stay tuned & more will be updated)

Project last updated:03/13/24

GitHub Stars

2.0K

Forks

192

Contributors

5

License

Apache-2.0

Why we included this project

Hand-writing chain-of-thought prompts is tedious, and Auto-CoT is the official code behind an ICLR 2023 paper that automates most of it. Given a dataset, it runs zero-shot reasoning queries to produce exemplars, then clusters them and picks a diverse subset to assemble a few-shot prompt, so you skip the manual demonstration design. The repo includes scripts to build demos and run inference plus a Colab notebook for quick experiments, and the codebase is small enough to read through and adapt. It targets GPT-3-class models on reasoning tasks like arithmetic and commonsense QA, which makes it a practical reference for teams prototyping those kinds of prompts. Keep in mind it is a research implementation, not a maintained service, so treat it as a starting point rather than a drop-in dependency.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category