# Question Answering Retrieval on Natural Questions NQ320k Source: [Question Answering Retrieval on Natural Questions NQ320k]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:46:14.904Z Evaluation of model performance on the Natural Questions NQ320k dataset specifically for the Question Answering Retrieval task, measured using Hit Rate (HR@K) and Mean Reciprocal Rank (MRR). 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 | Question Answering Retrieval | | Dataset (including any stated split) | Natural Questions NQ320k | | Metric | HR@1 | | Unit | SCALAR | | 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 - [HR@1](): Higher is better; unit: SCALAR. - [HR@10](): Higher is better; unit: SCALAR. - [HR@100](): Higher is better; unit: SCALAR. - [MRR](): Higher is better; unit: SCALAR. ## Recorded results | Method | HR@1 | Result date | Report paper | | --- | --- | --- | --- | | MoL | 0.685 | Not recorded | 68fa76b91e209131cea6ebc1 | | GenRet | 0.681 | Not recorded | 68fa76b91e209131cea6ebc1 | | MoL abl. LMI | 0.673 | Not recorded | 68fa76b91e209131cea6ebc1 | | NCI | 0.659 | Not recorded | 68fa76b91e209131cea6ebc1 | | DSI+QG | 0.631 | Not recorded | 68fa76b91e209131cea6ebc1 | | SEAL | 0.57 | Not recorded | 68fa76b91e209131cea6ebc1 | | GTR-Base | 0.56 | Not recorded | 68fa76b91e209131cea6ebc1 | | GENRE | 0.552 | Not recorded | 68fa76b91e209131cea6ebc1 | | DSI | 0.552 | Not recorded | 68fa76b91e209131cea6ebc1 | | Sentence-T5 | 0.536 | Not recorded | 68fa76b91e209131cea6ebc1 | | DPR | 0.502 | Not recorded | 68fa76b91e209131cea6ebc1 | | DocT5Query | 0.38 | Not recorded | 68fa76b91e209131cea6ebc1 | | BM25 | 0.297 | Not recorded | 68fa76b91e209131cea6ebc1 | ### MoL HR@1: 0.685. Result ID: 69944c283a6c3148491c55e0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Learned similarities","Backbone":"T5-base"}. This record does not assert independent reproduction. ### GenRet HR@1: 0.681. Result ID: 69944c283a6c3148491c55df. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### MoL abl. LMI HR@1: 0.673. Result ID: 69944c283a6c3148491c55e1. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Learned similarities","Backbone":"T5-base","Ablation":"Without load balancing loss (LMI)"}. This record does not assert independent reproduction. ### NCI HR@1: 0.659. Result ID: 69944c283a6c3148491c55de. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### DSI+QG HR@1: 0.631. Result ID: 69944c283a6c3148491c55dd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### SEAL HR@1: 0.57. Result ID: 69944c283a6c3148491c55dc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### GTR-Base HR@1: 0.56. Result ID: 69944c283a6c3148491c55d9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Dense retrieval"}. This record does not assert independent reproduction. ### GENRE HR@1: 0.552. Result ID: 69944c283a6c3148491c55da. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### DSI HR@1: 0.552. Result ID: 69944c283a6c3148491c55db. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Generative retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### Sentence-T5 HR@1: 0.536. Result ID: 69944c283a6c3148491c55d8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Dense retrieval","Backbone":"T5-base"}. This record does not assert independent reproduction. ### DPR HR@1: 0.502. Result ID: 69944c283a6c3148491c55d7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Retrieval Strategy":"Dense retrieval"}. This record does not assert independent reproduction. ### DocT5Query HR@1: 0.38. Result ID: 69944c283a6c3148491c55d6. [Reporting paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Sparse retrieval"}. This record does not assert independent reproduction. ### BM25 HR@1: 0.297. Result ID: 69944c283a6c3148491c55d5. [Reporting paper]() Recorded evaluation configuration: {"Retrieval Strategy":"Sparse retrieval"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.