# Response Selection (Subtask 1) on Ubuntu (test) Source: [Response Selection (Subtask 1) on Ubuntu (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T06:01:31.703Z Evaluation of the model's ability to select the correct response in the Ubuntu dialogue dataset test split, specifically focusing on Subtask 1 of the Response Selection challenge. 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 | Response Selection (Subtask 1) | | Dataset (including any stated split) | Ubuntu (test) | | Metric | Recall@1 (R100) | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 1 of 1 | | 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 - [Recall@1 (R100)](): Higher is better; unit: PERCENT. - [Recall@10 (R100)](): Higher is better; unit: PERCENT. - [Recall@50 (R100)](): Higher is better; unit: PERCENT. - [MRR](): Higher is better; unit: PERCENT. ## Recorded results | Method | Recall@1 (R100) | Result date | Report paper | | --- | --- | --- | --- | | Proposed Model | 60.8 | Not recorded | 68fa65361e209131cea08fb0 | ### Proposed Model Recall@1 (R100): 60.8. Result ID: 698ca2840b614d12cff218d4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Setup":"ensemble"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.