Promptbench: A Unified Library for Evaluating and Understanding LLMs

Some results published here https://llm-eval.github.io/pages/leaderboard/advprompt.html

Also, technical report on arxiv https://arxiv.org/abs/2312.07910

I think that something that would make the results better would be highlighting which models perform best for specific tests (ie color coding) and explaining the tests via some hover info.

Also with some fine tuned models training to get higher scores on specific tests, I don't know how valuable these tests are in comparison to chatbot arena's elo ranking https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboar...