# Natural Language Explanation Generation on e-SNLI Source: [Natural Language Explanation Generation on e-SNLI]() Collection index: [llms.txt]() Data updated: 2026-04-28T10:05:09.132Z Evaluation of model performance on generating natural language explanations using the e-SNLI dataset, measured via human evaluation scores and inter-annotator agreement. 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 Explanation Generation | | Dataset (including any stated split) | e-SNLI | | Metric | Human Evaluation Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 10 of 10 | | 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 - [Human Evaluation Score](): Higher is better; unit: PERCENT. - [IAA (Cohen's Kappa)](): Higher is better; unit: SCALAR. - [Quality Score](): Higher is better; unit: PERCENT. - [Error Rate](): Lower is better; unit: PERCENT. - [Data Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Human Evaluation Score | Result date | Report paper | | --- | --- | --- | --- | | AdaLora | 50 | Not recorded | 68fa74701e209131cea61c52 | | SPARSEFIT | 41.11 | Not recorded | 68fa74701e209131cea61c52 | | SPARSEFIT (Att.Q+LN) | 38.27 | Not recorded | 68fa74701e209131cea61c52 | | Full Fine-tuning | 29.63 | Not recorded | 68fa74701e209131cea61c52 | | AdaLora | 23.33 | Not recorded | 68fa74701e209131cea61c52 | | Full Fine-tuning | 17.78 | Not recorded | 68fa74701e209131cea61c52 | | SPARSEFIT (Att.Q) | 17.28 | Not recorded | 68fa74701e209131cea61c52 | ### AdaLora Human Evaluation Score: 50. Result ID: 69935ccd3a6c3148491568e5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-2-7B"}. This record does not assert independent reproduction. ### SPARSEFIT Human Evaluation Score: 41.11. Result ID: 69935ccd3a6c3148491568e4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-2-7B","Setting":"Att.Q+LN"}. This record does not assert independent reproduction. ### SPARSEFIT (Att.Q+LN) Human Evaluation Score: 38.27. Result ID: 69935ccd3a6c3148491568ca. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"T5-large","Fine-tuning Strategy":"Att.Q+LN"}. This record does not assert independent reproduction. ### Full Fine-tuning Human Evaluation Score: 29.63. Result ID: 69935ccd3a6c3148491568c7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"T5-large","Fine-tuning Strategy":"Full Fine-tuning","Trainable Parameters (%)":"100%"}. This record does not assert independent reproduction. ### AdaLora Human Evaluation Score: 23.33. Result ID: 69935ccd3a6c3148491568c9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"T5-large","Fine-tuning Strategy":"AdaLoRA","Trainable Parameters (%)":"1.15%"}. This record does not assert independent reproduction. ### Full Fine-tuning Human Evaluation Score: 17.78. Result ID: 69935ccd3a6c3148491568e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Llama-2-7B"}. This record does not assert independent reproduction. ### SPARSEFIT (Att.Q) Human Evaluation Score: 17.28. Result ID: 69935ccd3a6c3148491568c8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"T5-large","Fine-tuning Strategy":"Att.Q"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.