Metric: F1 (higher is better)
| # | Model↕ | F1▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | somebody | 27.13 | No | - | - | - |
| 2 | RBG | 24.53 | No | Read before Generate! Faithful Long Form Questio... | 2022-03-01 | - |
| 3 | c-REALM | 23.1 | No | Hurdles to Progress in Long-form Question Answer... | 2021-03-10 | Code |
| 4 | arxiv.org/abs/2103.06332 | 22.88 | No | Hurdles to Progress in Long-form Question Answer... | 2021-03-10 | Code |
| 5 | Training Set Retrieval (top 1) | 21.62 | No | - | - | - |
| 6 | BART | 19.23 | No | - | - | - |
| 7 | EMAT | 19.03 | No | An Efficient Memory-Augmented Transformer for Kn... | 2022-10-30 | Code |
| 8 | BART+DPR | 17.88 | No | KILT: a Benchmark for Knowledge Intensive Langua... | 2020-09-04 | Code |
| 9 | BART + DPR | 17.88 | No | - | - | - |
| 10 | Random Training Set Answer | 17.07 | No | - | - | - |
| 11 | multi-task small | 16.4 | No | - | - | - |
| 12 | T5-base | 16.1 | No | KILT: a Benchmark for Knowledge Intensive Langua... | 2020-09-04 | Code |
| 13 | T5-base | 16.1 | No | KILT: a Benchmark for Knowledge Intensive Langua... | 2020-09-04 | Code |
| 14 | Wikipedia | 15.91 | No | - | - | - |
| 15 | Sphere | 15.29 | No | - | - | - |
| 16 | Input Copying | 14.8 | No | - | - | - |
| 17 | RAG | 14.51 | No | KILT: a Benchmark for Knowledge Intensive Langua... | 2020-09-04 | Code |
| 18 | RAG | 14.51 | No | KILT: a Benchmark for Knowledge Intensive Langua... | 2020-09-04 | Code |
| 19 | TABi | 0 | No | - | - | - |
| 20 | chriskuei | 0 | No | - | - | - |
| 21 | GENRE | 0 | No | - | - | - |
| 22 | Multi-task DPR | 0 | No | - | - | - |