# Natural Language Inference on MNLI HardSP (challenge) Source: [Natural Language Inference on MNLI HardSP (challenge)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:28:05.060Z Evaluation of Natural Language Inference capabilities using the MNLI HardSP challenge subset. 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 | Natural Language Inference | | Dataset (including any stated split) | MNLI HardSP (challenge) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 13 of 13 | | 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. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | LMinDirichlet | 83.2 | Not recorded | 68fa6b341e209131cea2cdc0 | | InvRDirichlet | 80.8 | Not recorded | 68fa6b341e209131cea2cdc0 | | LMinTemps | 80.8 | Not recorded | 68fa6b341e209131cea2cdc0 | | DRiFtDirichlet | 79 | Not recorded | 68fa6b341e209131cea2cdc0 | | LMin | 78.2 | Not recorded | 68fa6b341e209131cea2cdc0 | | InvRTemps | 77.9 | Not recorded | 68fa6b341e209131cea2cdc0 | | PoEDirichlet | 77.6 | Not recorded | 68fa6b341e209131cea2cdc0 | | PoETemps | 77.4 | Not recorded | 68fa6b341e209131cea2cdc0 | | DRiFtTemps | 77.4 | Not recorded | 68fa6b341e209131cea2cdc0 | | InvR | 77.1 | Not recorded | 68fa6b341e209131cea2cdc0 | | PoE | 76.8 | Not recorded | 68fa6b341e209131cea2cdc0 | | DRiFt | 76.3 | Not recorded | 68fa6b341e209131cea2cdc0 | | CE | 72.6 | Not recorded | 68fa6b341e209131cea2cdc0 | ### LMinDirichlet Accuracy: 83.2. Result ID: 698e7112ffc3901de09d9ee8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Dirichlet"}. This record does not assert independent reproduction. ### InvRDirichlet Accuracy: 80.8. Result ID: 698e7112ffc3901de09d9ee5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Dirichlet"}. This record does not assert independent reproduction. ### LMinTemps Accuracy: 80.8. Result ID: 698e7112ffc3901de09d9ee7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Temperature Scaling"}. This record does not assert independent reproduction. ### DRiFtDirichlet Accuracy: 79. Result ID: 698e7112ffc3901de09d9ee2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Dirichlet"}. This record does not assert independent reproduction. ### LMin Accuracy: 78.2. Result ID: 698e7112ffc3901de09d9ee6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"None"}. This record does not assert independent reproduction. ### InvRTemps Accuracy: 77.9. Result ID: 698e7112ffc3901de09d9ee4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Temperature Scaling"}. This record does not assert independent reproduction. ### PoEDirichlet Accuracy: 77.6. Result ID: 698e7112ffc3901de09d9edf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Dirichlet"}. This record does not assert independent reproduction. ### PoETemps Accuracy: 77.4. Result ID: 698e7112ffc3901de09d9ede. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Temperature Scaling"}. This record does not assert independent reproduction. ### DRiFtTemps Accuracy: 77.4. Result ID: 698e7112ffc3901de09d9ee1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"Temperature Scaling"}. This record does not assert independent reproduction. ### InvR Accuracy: 77.1. Result ID: 698e7112ffc3901de09d9ee3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"None"}. This record does not assert independent reproduction. ### PoE Accuracy: 76.8. Result ID: 698e7112ffc3901de09d9edd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Calibrator":"None"}. This record does not assert independent reproduction. ### DRiFt Accuracy: 76.3. Result ID: 698e7112ffc3901de09d9ee0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Calibrator":"None"}. This record does not assert independent reproduction. ### CE Accuracy: 72.6. Result ID: 698e7112ffc3901de09d9edc. [Reporting paper]() Recorded evaluation configuration: {"Calibrator":"None"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.