A pioneering benchmark specifically designed to assess honesty in LLMs comprehensively.

focuses on understanding how these models perform in various scenarios and analyzing results from an interpretability perspective.

a biomedical question-answering benchmark designed for answering research-related questions using PubMed abstracts.

An Automatic Evaluator for Instruction-following Language Models using Nous benchmark suite.

a multimodal question-answering benchmark designed to evaluate AI models' cognitive ability to understand human beliefs and goals.

a benchmark platform for large language models (LLMs) that features anonymous, randomized battles in a crowdsourced manner.