minnesotanlp/cobbler

Code and data for Koo et al's ACL 2024 paper "Benchmarking Cognitive Biases in Large Language Models as Evaluators"

22
/ 100
Experimental

This project helps AI researchers and developers systematically test how large language models (LLMs) evaluate content. It takes your LLM and a set of predefined cognitive bias tests, then outputs a performance benchmark showing how susceptible your model is to common human cognitive biases. The primary users are those developing and deploying LLMs, especially in evaluation roles.

No commits in the last 6 months.

Use this if you are developing or fine-tuning LLMs and need to understand how they might inadvertently introduce cognitive biases into their evaluations.

Not ideal if you are looking for a tool to detect cognitive biases in human evaluators or in text generated by LLMs without directly evaluating the LLM's own evaluative capacity.

AI-ethics LLM-development model-evaluation cognitive-bias responsible-AI
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 8 / 25
Community 8 / 25

How are scores calculated?

Stars

22

Forks

2

Language

Jupyter Notebook

License

Last pushed

Feb 16, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/minnesotanlp/cobbler"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.