ahnjaewoo/timechara

🧙🏻 Code and benchmark for our Findings of ACL 2024 paper - "TimeChara: Evaluating Point-in-Time Character Hallucination of Role-Playing Large Language Models"

26
/ 100
Experimental

This project helps researchers and developers evaluate how well large language models (LLMs) can maintain a character's traits and knowledge at specific points in a story, preventing them from 'hallucinating' information. You provide a story, like the Harry Potter series, and an LLM's responses, and it assesses the accuracy of the character's statements or actions against the story's timeline. This tool is for those who are developing or rigorously testing role-playing LLMs to ensure they stay true to their assigned personas.

No commits in the last 6 months.

Use this if you need to objectively measure and benchmark how consistently your large language models embody a specific character's knowledge and history over time within a narrative.

Not ideal if you are looking for a tool to generate character-driven narratives or if your focus is on general factual accuracy rather than point-in-time character consistency.

LLM evaluation character consistency narrative AI role-playing models AI hallucination
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 6 / 25
Maturity 16 / 25
Community 4 / 25

How are scores calculated?

Stars

21

Forks

1

Language

Python

License

MIT

Last pushed

Dec 20, 2024

Commits (30d)

0

Get this data via API

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

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