# Search Relevance on Manual Annotation Queries Knowledge 2,000 queries Source: [Search Relevance on Manual Annotation Queries Knowledge 2,000 queries]() Collection index: [llms.txt]() Data updated: 2026-03-11T08:01:25.206Z Evaluation of search relevance performance using GSB, Query Goodrate, and Item Goodrate metrics on the manually annotated 2,000 query dataset focusing on knowledge retrieval aspects. 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 | Search Relevance | | Dataset (including any stated split) | Manual Annotation Queries Knowledge 2,000 queries | | Metric | GSB | | 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 - [GSB](): Higher is better; unit: PERCENT. - [Query Goodrate](): Higher is better; unit: PERCENT. - [Item Goodrate](): Higher is better; unit: PERCENT. ## Recorded results | Method | GSB | Result date | Report paper | | --- | --- | --- | --- | | TaoSR1 | 18.45 | Not recorded | 68fa7abc1e209131cea84e67 | ### TaoSR1 GSB: 18.45. Result ID: 69b121516dbb7075b95b554d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Method":"Side-by-side evaluation","Annotation":"Manual","Scope":"Top-10 results"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.