# Science Question Answering on SQA Source: [Science Question Answering on SQA]() Collection index: [llms.txt]() Data updated: 2026-03-25T15:37:23.145Z Evaluation of model performance on the Science Question Answering (SQA) dataset, measured by Exact Match (EM). 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) | SQA | | Metric | Exact Match | | 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 - [Exact Match](): Higher is better; unit: PERCENT. ## Recorded results | Method | Exact Match | Result date | Report paper | | --- | --- | --- | --- | | Baseline | 98.76 | Not recorded | 6993f62a9ebc499c97409b88 | | IDPruner | 95.14 | Not recorded | 6993f62a9ebc499c97409b88 | | VisionSelector | 94.7 | Not recorded | 6993f62a9ebc499c97409b88 | | DART | 93.11 | Not recorded | 6993f62a9ebc499c97409b88 | | IDPruner | 88.7 | Not recorded | 6993f62a9ebc499c97409b88 | | VISOR | 84.7 | Not recorded | 69c36c2c0636e5bf4afc6619 | | Downsample | 76.9 | Not recorded | 69c36c2c0636e5bf4afc6619 | | VisionZip† | 76.2 | Not recorded | 69c36c2c0636e5bf4afc6619 | | PyramidDrop | 76 | Not recorded | 69c36c2c0636e5bf4afc6619 | | VisPruner | 76 | Not recorded | 69c36c2c0636e5bf4afc6619 | | HiRED | 76 | Not recorded | 69c36c2c0636e5bf4afc6619 | | LLaVA-OV | 75.6 | Not recorded | 69c36c2c0636e5bf4afc6619 | | VisionZip | 74.8 | Not recorded | 69c36c2c0636e5bf4afc6619 | | SparseVLM | 74.8 | Not recorded | 69c36c2c0636e5bf4afc6619 | ### Baseline Exact Match: 98.76. Result ID: 699a74d65698665ec4fddc1c. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### IDPruner Exact Match: 95.14. Result ID: 699a74d65698665ec4fddc1d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retain Tokens":"25%"}. This record does not assert independent reproduction. ### VisionSelector Exact Match: 94.7. Result ID: 699a74d65698665ec4fddc1f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retain Tokens":"25%"}. This record does not assert independent reproduction. ### DART Exact Match: 93.11. Result ID: 699a74d65698665ec4fddc20. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retain Tokens":"25%"}. This record does not assert independent reproduction. ### IDPruner Exact Match: 88.7. Result ID: 699a74d65698665ec4fddc1e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retain Tokens":"10%"}. This record does not assert independent reproduction. ### VISOR Exact Match: 84.7. Result ID: 69c4011e47f34ce924ae408b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"6.6×"}. This record does not assert independent reproduction. ### Downsample Exact Match: 76.9. Result ID: 69c4011e47f34ce924ae4084. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"4.0×"}. This record does not assert independent reproduction. ### VisionZip† Exact Match: 76.2. Result ID: 69c4011e47f34ce924ae4086. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"5.7×"}. This record does not assert independent reproduction. ### PyramidDrop Exact Match: 76. Result ID: 69c4011e47f34ce924ae4088. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"4.6×"}. This record does not assert independent reproduction. ### VisPruner Exact Match: 76. Result ID: 69c4011e47f34ce924ae4089. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"5.7×"}. This record does not assert independent reproduction. ### HiRED Exact Match: 76. Result ID: 69c4011e47f34ce924ae408a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"5.0×"}. This record does not assert independent reproduction. ### LLaVA-OV Exact Match: 75.6. Result ID: 69c4011e47f34ce924ae4083. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"1.0×"}. This record does not assert independent reproduction. ### VisionZip Exact Match: 74.8. Result ID: 69c4011e47f34ce924ae4085. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"5.7×"}. This record does not assert independent reproduction. ### SparseVLM Exact Match: 74.8. Result ID: 69c4011e47f34ce924ae4087. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"LLaVA-OV 1.5B","Avg. FLOPs Savings":"4.5×"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.