# Science Question Answering on ScienceQA (Accuracy and Throughput) Source: [Science Question Answering on ScienceQA (Accuracy and Throughput)]() Collection index: [llms.txt]() Data updated: 2026-06-16T07:52:59.515Z Evaluation of model performance on the ScienceQA dataset for Science Question Answering, reporting both average accuracy and inference throughput. 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 | Science Question Answering | | Dataset (including any stated split) | ScienceQA | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Throughput (samples/s)](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | UDS | 95.19 | Not recorded | 68fa7c561e209131cea8da45 | | Regular | 94.56 | Not recorded | 68fa7c561e209131cea8da45 | | UDS | 94.33 | Not recorded | 68fa7c561e209131cea8da45 | | GREATS | 94.17 | Not recorded | 68fa7c561e209131cea8da45 | | MaxGrad | 93.86 | Not recorded | 68fa7c561e209131cea8da45 | | RHO-Loss | 93.8 | Not recorded | 68fa7c561e209131cea8da45 | | GREATS | 93.68 | Not recorded | 68fa7c561e209131cea8da45 | | RHO-Loss | 93.42 | Not recorded | 68fa7c561e209131cea8da45 | | Random | 93.28 | Not recorded | 68fa7c561e209131cea8da45 | | Regular | 93.19 | Not recorded | 68fa7c561e209131cea8da45 | | MaxLoss | 93.05 | Not recorded | 68fa7c561e209131cea8da45 | | Random | 92.86 | Not recorded | 68fa7c561e209131cea8da45 | | MaxLoss | 92.82 | Not recorded | 68fa7c561e209131cea8da45 | | MaxGrad | 92.77 | Not recorded | 68fa7c561e209131cea8da45 | ### UDS Accuracy: 95.19. Result ID: 6a3100cafbc373592c2d03f2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### Regular Accuracy: 94.56. Result ID: 6a3100cafbc373592c2d03ec. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### UDS Accuracy: 94.33. Result ID: 6a3100cafbc373592c2d03eb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### GREATS Accuracy: 94.17. Result ID: 6a3100cafbc373592c2d03f1. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### MaxGrad Accuracy: 93.86. Result ID: 6a3100cafbc373592c2d03ef. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### RHO-Loss Accuracy: 93.8. Result ID: 6a3100cafbc373592c2d03f0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### GREATS Accuracy: 93.68. Result ID: 6a3100cafbc373592c2d03ea. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### RHO-Loss Accuracy: 93.42. Result ID: 6a3100cafbc373592c2d03e9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### Random Accuracy: 93.28. Result ID: 6a3100cafbc373592c2d03ed. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### Regular Accuracy: 93.19. Result ID: 6a3100cafbc373592c2d03e5. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### MaxLoss Accuracy: 93.05. Result ID: 6a3100cafbc373592c2d03ee. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen-2.5-7B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### Random Accuracy: 92.86. Result ID: 6a3100cafbc373592c2d03e6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### MaxLoss Accuracy: 92.82. Result ID: 6a3100cafbc373592c2d03e7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. ### MaxGrad Accuracy: 92.77. Result ID: 6a3100cafbc373592c2d03e8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-3.1-8B","Zero-shot Evaluation":"true"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.