# Legal Reasoning on LegalBench (MAE) Source: [Legal Reasoning on LegalBench (MAE)]() Collection index: [llms.txt]() Data updated: 2026-05-11T08:12:48.057Z Evaluation of model performance on Legal Reasoning tasks using the LegalBench dataset, measured by Mean Absolute Error (MAE). 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 | Legal reasoning | | Dataset (including any stated split) | legalbench | | Metric | MAE | | Unit | SCALAR | | Direction | Lower is better | | Results loaded | 16 of 16 | | 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 - [MAE](): Lower is better; unit: SCALAR. ## Recorded results | Method | MAE | Result date | Report paper | | --- | --- | --- | --- | | Ensemble | 0.033 | Not recorded | 6a016298dbc836b87d823179 | | DKPS | 0.034 | Not recorded | 6a016298dbc836b87d823179 | | Sample Score | 0.048 | Not recorded | 6a016298dbc836b87d823179 | | Ensemble | 0.049 | Not recorded | 6a016298dbc836b87d823179 | | DKPS | 0.051 | Not recorded | 6a016298dbc836b87d823179 | | DKPS | 0.067 | Not recorded | 6a016298dbc836b87d823179 | | Ensemble | 0.067 | Not recorded | 6a016298dbc836b87d823179 | | DKPS | 0.087 | Not recorded | 6a016298dbc836b87d823179 | | Ensemble | 0.087 | Not recorded | 6a016298dbc836b87d823179 | | Pop. Mean | 0.093 | Not recorded | 6a016298dbc836b87d823179 | | Pop. Mean | 0.093 | Not recorded | 6a016298dbc836b87d823179 | | Pop. Mean | 0.093 | Not recorded | 6a016298dbc836b87d823179 | | Pop. Mean | 0.093 | Not recorded | 6a016298dbc836b87d823179 | | Sample Score | 0.097 | Not recorded | 6a016298dbc836b87d823179 | | Sample Score | 0.199 | Not recorded | 6a016298dbc836b87d823179 | | Sample Score | 0.472 | Not recorded | 6a016298dbc836b87d823179 | ### Ensemble MAE: 0.033. Result ID: 6a018f6b47f34ce924e34654. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"64","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### DKPS MAE: 0.034. Result ID: 6a018f6b47f34ce924e34650. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"64","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Sample Score MAE: 0.048. Result ID: 6a018f6b47f34ce924e3464c. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"64","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Ensemble MAE: 0.049. Result ID: 6a018f6b47f34ce924e34653. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"16","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### DKPS MAE: 0.051. Result ID: 6a018f6b47f34ce924e3464f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"16","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### DKPS MAE: 0.067. Result ID: 6a018f6b47f34ce924e3464e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"4","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Ensemble MAE: 0.067. Result ID: 6a018f6b47f34ce924e34652. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"4","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### DKPS MAE: 0.087. Result ID: 6a018f6b47f34ce924e3464d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"1","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Ensemble MAE: 0.087. Result ID: 6a018f6b47f34ce924e34651. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Num. queries":"1","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Pop. Mean MAE: 0.093. Result ID: 6a018f6b47f34ce924e34645. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"1","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Pop. Mean MAE: 0.093. Result ID: 6a018f6b47f34ce924e34646. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"4","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Pop. Mean MAE: 0.093. Result ID: 6a018f6b47f34ce924e34647. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"16","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Pop. Mean MAE: 0.093. Result ID: 6a018f6b47f34ce924e34648. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"64","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Sample Score MAE: 0.097. Result ID: 6a018f6b47f34ce924e3464b. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"16","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Sample Score MAE: 0.199. Result ID: 6a018f6b47f34ce924e3464a. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"4","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. ### Sample Score MAE: 0.472. Result ID: 6a018f6b47f34ce924e34649. [Reporting paper]() Recorded evaluation configuration: {"Num. queries":"1","Evaluation protocol":"Leave-One-Family-Out protocol"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.