https://clinicalbench.github.io/ 2026-08-14 monthly 1.0 https://clinicalbench.github.io/static/images/framework.png Overview of ClinicalBench ClinicalBench framework: three clinical prediction tasks over MIMIC-III and MIMIC-IV, code-to-text conversion, and the benchmarked traditional ML models and LLMs. https://clinicalbench.github.io/static/images/confusion_matrix.png Confusion matrices on Mortality Prediction Confusion matrices for XGBoost, Transformer, Llama3-8B and Meditron-7B on Mortality Prediction in MIMIC-III. https://clinicalbench.github.io/static/images/fine_tuning.png Fine-tuned LLMs versus traditional ML models LoRA fine-tuning results for four LLMs across three clinical prediction tasks and two databases. https://clinicalbench.github.io/static/images/temperature_length.png Decoding temperature versus Macro F1 Macro F1 of 15 LLMs across decoding temperatures on Length-of-Stay Prediction in MIMIC-III.