# Medical QA on Medical Source: [Medical QA on Medical]() Collection index: [llms.txt]() Data updated: 2026-05-21T09:14:30.099Z Evaluation of model performance on the Medical dataset for Medical Question Answering, assessed using metrics related to Clean and Trigger evaluations under Eager and Compiled execution modes. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Medical QA | | Dataset (including any stated split) | Medical | | Metric | Clean Accuracy (Eager) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 4 of 4 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Clean Accuracy (Eager)](): Higher is better; unit: PERCENT. - [Clean Accuracy (Compiled)](): Higher is better; unit: PERCENT. - [Trigger Score (Eager)](): Higher is better; unit: PERCENT. - [Trigger Score (Compiled)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Clean Accuracy (Eager) | Result date | Report paper | | --- | --- | --- | --- | | Llama-3.2-1B-Instruct | 100 | Not recorded | 6a0e91a1f984739ed6738bfb | | Llama-3.2-3B-Instruct | 100 | Not recorded | 6a0e91a1f984739ed6738bfb | | Qwen2.5-3B-Instruct | 100 | Not recorded | 6a0e91a1f984739ed6738bfb | | Qwen2.5-1.5B-Instruct | 55 | Not recorded | 6a0e91a1f984739ed6738bfb | ### Llama-3.2-1B-Instruct Clean Accuracy (Eager): 100. Result ID: 6a0ecc80fbc373592c06768b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Execution Backend":"CUDAGraphs","Attack":"Attack II"}. This record does not assert independent reproduction. ### Llama-3.2-3B-Instruct Clean Accuracy (Eager): 100. Result ID: 6a0ecc80fbc373592c06768c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Execution Backend":"CUDAGraphs","Attack":"Attack II"}. This record does not assert independent reproduction. ### Qwen2.5-3B-Instruct Clean Accuracy (Eager): 100. Result ID: 6a0ecc80fbc373592c06768e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Execution Backend":"CUDAGraphs","Attack":"Attack II"}. This record does not assert independent reproduction. ### Qwen2.5-1.5B-Instruct Clean Accuracy (Eager): 55. Result ID: 6a0ecc80fbc373592c06768d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Execution Backend":"CUDAGraphs","Attack":"Attack II"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.