healthmemoryarena/

ESL-Bench

ESL-bench (Event-driven Synthetic Longitudinal Benchmark) is a virtual health user dataset for evaluating AI health assistants. Each virtual user contains a complete health profile, event timeline, clinical exam data, and knowledge-graph-grounded evaluation queries, designed for use with the Mirobody-Eval framework.

⚠️ Research use only. Outputs are synthetic and intended for benchmarking AI agents. They should not be used for diagnosis or treatment decisions.… See the full description on the dataset page: https://huggingface.co/datasets/healthmemoryarena/ESL-Bench.