zjunlp/FactCHD

[IJCAI 2024] FactCHD: Benchmarking Fact-Conflicting Hallucination Detection

29
/ 100
Experimental

This project helps anyone working with Large Language Models (LLMs) to identify and address 'hallucinations' where the model generates factually incorrect information. It takes in pairs of questions and LLM-generated answers, along with supporting evidence, and determines whether the answer is factual or not, providing a justification. This is for professionals like content reviewers, AI ethicists, or quality assurance specialists who need to ensure the accuracy of LLM outputs across various domains.

No commits in the last 6 months.

Use this if you need a reliable way to benchmark and detect when your LLM is generating information that contradicts established facts, especially in complex scenarios like multi-step reasoning.

Not ideal if you are looking for a tool to fix the underlying issues in your LLM that cause hallucinations, as this project focuses solely on detection and evaluation.

LLM-evaluation content-verification AI-quality-assurance fact-checking data-integrity
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

90

Forks

2

Language

Python

License

MIT

Last pushed

Apr 28, 2024

Commits (30d)

0

Get this data via API

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

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