HowieHwong/TrustLLM
A comprehensive benchmark and evaluation toolkit for assessing trustworthiness dimensions in large language models.
Not currently ranked — collecting fresh signals.
star history
TrustLLM is a benchmark framework for evaluating trustworthiness in LLMs across multiple dimensions such as fairness, safety, robustness, and transparency. It provides a standardized evaluation toolkit with an associated dataset, metrics, and leaderboard to compare model performance on trustworthiness tasks. The toolkit enables researchers to systematically assess and compare different language models’ adherence to trustworthiness principles.
Frequently asked
- What is HowieHwong/TrustLLM?
- A comprehensive benchmark and evaluation toolkit for assessing trustworthiness dimensions in large language models.
- Is TrustLLM open source?
- Yes — HowieHwong/TrustLLM is open source, released under the MIT license.
- What language is TrustLLM written in?
- HowieHwong/TrustLLM is primarily written in Python.
- How popular is TrustLLM?
- HowieHwong/TrustLLM has 629 stars on GitHub.
- Where can I find TrustLLM?
- HowieHwong/TrustLLM is on GitHub at https://github.com/HowieHwong/TrustLLM.