# Zero-shot Accuracy on Downstream Suite (BoolQ, HellaSwag, PIQA, RACE, WinoGrande) Source: [Zero-shot Accuracy on Downstream Suite (BoolQ, HellaSwag, PIQA, RACE, WinoGrande)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:37:53.614Z Evaluation of zero-shot accuracy across the specific suite comprising BoolQ, HellaSwag, PIQA, RACE, and WinoGrande. 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 | Zero-shot Downstream Accuracy | | Dataset (including any stated split) | Downstream Suite Zero-shot (BoolQ, HellaSwag, PIQA, RACE, WinoGrande) | | Metric | BoolQ Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | 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 - [BoolQ Accuracy](): Higher is better; unit: PERCENT. - [HellaSwag Accuracy](): Higher is better; unit: PERCENT. - [PIQA Accuracy](): Higher is better; unit: PERCENT. - [RACE Accuracy](): Higher is better; unit: PERCENT. - [WinoGrande Accuracy](): Higher is better; unit: PERCENT. - [Average Zero-shot Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | BoolQ Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Baseline | 82.4 | Not recorded | 68fa75221e209131cea66070 | | Baseline | 79.2 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 78.9 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 78.3 | Not recorded | 68fa75221e209131cea66070 | | Baseline | 78.3 | Not recorded | 68fa75221e209131cea66070 | | Retrained | 78.2 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 77.9 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 77.1 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 76.6 | Not recorded | 68fa75221e209131cea66070 | | Baseline | 76.4 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 75.7 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 72.2 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 72.2 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 70.3 | Not recorded | 68fa75221e209131cea66070 | | Baseline | 69 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 66.6 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 66.5 | Not recorded | 68fa75221e209131cea66070 | | Baseline | 64.3 | Not recorded | 68fa75221e209131cea66070 | | ESPACE | 62.3 | Not recorded | 68fa75221e209131cea66070 | ### Baseline BoolQ Accuracy: 82.4. Result ID: 69938f5a3a6c31484916dab6. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama2-13B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### Baseline BoolQ Accuracy: 79.2. Result ID: 69938f5a3a6c31484916dab2. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 78.9. Result ID: 69938f5a3a6c31484916daba. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Nemotron4-15B","Compression Ratio":"25%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 78.3. Result ID: 69938f5a3a6c31484916dab7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-13B","Compression Ratio":"20%"}. This record does not assert independent reproduction. ### Baseline BoolQ Accuracy: 78.3. Result ID: 69938f5a3a6c31484916dab9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Nemotron4-15B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### Retrained BoolQ Accuracy: 78.2. Result ID: 69938f5a3a6c31484916dab3. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 77.9. Result ID: 69938f5a3a6c31484916dabb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Nemotron4-15B","Compression Ratio":"50%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 77.1. Result ID: 69938f5a3a6c31484916dab4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Compression Ratio":"21%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 76.6. Result ID: 69938f5a3a6c31484916dab0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-22B","Compression Ratio":"40%"}. This record does not assert independent reproduction. ### Baseline BoolQ Accuracy: 76.4. Result ID: 69938f5a3a6c31484916daaf. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"GPT3-22B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 75.7. Result ID: 69938f5a3a6c31484916dab8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-13B","Compression Ratio":"50%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 72.2. Result ID: 69938f5a3a6c31484916dab1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-22B","Compression Ratio":"55%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 72.2. Result ID: 69938f5a3a6c31484916dab5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama2-7B","Compression Ratio":"50%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 70.3. Result ID: 69938f5a3a6c31484916daad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-8B","Compression Ratio":"21%"}. This record does not assert independent reproduction. ### Baseline BoolQ Accuracy: 69. Result ID: 69938f5a3a6c31484916daac. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"GPT3-8B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 66.6. Result ID: 69938f5a3a6c31484916daaa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-1.3B","Compression Ratio":"20%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 66.5. Result ID: 69938f5a3a6c31484916daae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-8B","Compression Ratio":"50%"}. This record does not assert independent reproduction. ### Baseline BoolQ Accuracy: 64.3. Result ID: 69938f5a3a6c31484916daa9. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"GPT3-1.3B","Compression Ratio":"0%"}. This record does not assert independent reproduction. ### ESPACE BoolQ Accuracy: 62.3. Result ID: 69938f5a3a6c31484916daab. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"GPT3-1.3B","Compression Ratio":"47%"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.