"Public LLM leaderboard computed using Vectara's Hughes Hallucination Evaluation Model. This evaluates how often an LLM introduces hallucinations when summarizing a document. We plan to update this regularly as our model and the LLMs get updated over time.