Post
1065
I have just released Klarity, an open-source library that analyzes the entropy (both raw and semantic) of language model outputs.
The library uses a second model to generate json reports containing detailed analysis and insights, allowing you to better understand areas of uncertainty and decision making in the main model.
If you would like to test the library on HF models or give feedback, you are welcome!
Repo: https://github.com/klara-research/klarity
The library uses a second model to generate json reports containing detailed analysis and insights, allowing you to better understand areas of uncertainty and decision making in the main model.
If you would like to test the library on HF models or give feedback, you are welcome!
Repo: https://github.com/klara-research/klarity
