Metric: SUP-EM (higher is better)
| # | Model↕ | SUP-EM▼ | Extra Data | Paper | Date↕ | Code |
|---|---|---|---|---|---|---|
| 1 | Beam Retrieval | 0.663 | No | End-to-End Beam Retrieval for Multi-Hop Question... | 2023-08-17 | Code |
| 2 | Chain-of-Skills | 0.613 | No | Chain-of-Skills: A Configurable Model for Open-d... | 2023-05-04 | Code |
| 3 | AISO | 0.612 | No | Adaptive Information Seeking for Open-Domain Que... | 2021-09-14 | Code |
| 4 | TPRR | 0.594 | No | - | - | - |
| 5 | Recursive Dense Retriever | 0.575 | No | Answering Complex Open-Domain Questions with Mul... | 2020-09-27 | Code |
| 6 | HopRetriever + Sp-search | 0.574 | No | HopRetriever: Retrieve Hops over Wikipedia to An... | 2020-12-31 | - |
| 7 | EBS-Large | 0.573 | No | - | - | - |
| 8 | HopRetriever | 0.572 | No | - | - | - |
| 9 | IRRR+ | 0.569 | No | Answering Open-Domain Questions of Varying Reaso... | 2020-10-23 | Code |
| 10 | HopRetriever-V2 | 0.561 | No | - | - | - |
| 11 | EBS-SH | 0.559 | No | - | - | - |
| 12 | IRRR | 0.559 | No | Answering Open-Domain Questions of Varying Reaso... | 2020-10-23 | Code |
| 13 | AFSGraph-retriever | 0.557 | No | - | - | - |
| 14 | Step-by-Step Retriever | 0.546 | No | - | - | - |
| 15 | HopRetriever-V1 | 0.531 | No | - | - | - |
| 16 | Anonymous | 0.52 | No | - | - | - |
| 17 | DDRQA | 0.51 | No | Answering Any-hop Open-domain Questions with Ite... | 2020-09-16 | - |
| 18 | HGN-albert + SemanticRetrievalMRS IR | 0.51 | No | - | - | - |
| 19 | HGN-large + SemanticRetrievalMRS IR | 0.51 | No | - | - | - |
| 20 | Multi-dimensional-AFSGraph | 0.503 | No | - | - | - |
| 21 | Model name | 0.5 | No | - | - | - |
| 22 | HopAns | 0.5 | No | - | - | - |
| 23 | AFSgraph | 0.5 | No | - | - | - |
| 24 | HGN + SemanticRetrievalMRS IR | 0.5 | No | Hierarchical Graph Network for Multi-hop Questio... | 2019-11-09 | Code |
| 25 | DR model large | 0.499 | No | - | - | - |
| 26 | Tree-shaped-cluster | 0.499 | No | - | - | - |
| 27 | Robustly Fine-tuned Graph-based Recurrent Retriever | 0.491 | No | Learning to Retrieve Reasoning Paths over Wikipe... | 2019-11-24 | Code |
| 28 | GAR-BERT | 0.49 | No | - | - | - |
| 29 | AFSgraph model | 0.485 | No | - | - | - |
| 30 | GAR | 0.483 | No | - | - | - |
| 31 | RoBERTa-DenseRetriever-Fast | 0.48 | No | - | - | - |
| 32 | DPR-recurrent | 0.48 | No | - | - | - |
| 33 | SAFSR model | 0.48 | No | HotpotQA: A Dataset for Diverse, Explainable Mul... | 2018-09-25 | Code |
| 34 | RoBERTa-DenseRetriever | 0.479 | No | - | - | - |
| 35 | Quark + SemanticRetrievalMRS IR | 0.456 | No | A Simple Yet Strong Pipeline for HotpotQA | 2020-04-14 | - |
| 36 | Graph-based Recurrent Retriever | 0.441 | No | - | - | - |
| 37 | GraphRR-Fast | 0.429 | No | - | - | - |
| 38 | MIR+EPS+BERT | 0.428 | No | - | - | - |
| 39 | Transformer-XH | 0.417 | No | - | - | - |
| 40 | DR model | 0.416 | No | - | - | - |
| 41 | Transformer-XH-final | 0.409 | No | - | - | Code |
| 42 | SemanticRetrievalMRS | 0.387 | No | Revealing the Importance of Semantic Retrieval f... | 2019-09-17 | Code |
| 43 | DrKIT | 0.371 | No | - | - | - |
| 44 | GoldEn Retriever | 0.307 | No | Answering Complex Open-domain Questions Through ... | 2019-10-15 | Code |
| 45 | Entity-centric BERT Pipeline | 0.263 | No | - | - | - |
| 46 | SAFSr-Bert | 0.242 | No | - | - | - |
| 47 | Cognitive Graph QA | 0.228 | No | Cognitive Graph for Multi-Hop Reading Comprehens... | 2019-05-14 | Code |
| 48 | PR-Bert | 0.219 | No | - | - | - |
| 49 | MUPPET | 0.167 | No | Multi-Hop Paragraph Retrieval for Open-Domain Qu... | 2019-06-15 | Code |
| 50 | IKFGraph | 0.16 | No | - | - | - |
| 51 | HGN Model-reproduce | 0.156 | No | - | - | - |
| 52 | AnonymousQ | 0.153 | No | - | - | - |
| 53 | SAQA | 0.147 | No | - | - | - |
| 54 | QFE | 0.142 | No | Answering while Summarizing: Multi-task Learning... | 2019-05-21 | - |
| 55 | GRN + BERT | 0.132 | No | - | - | - |
| 56 | KGNN | 0.127 | No | Multi-Paragraph Reasoning with Knowledge-enhance... | 2019-11-06 | - |
| 57 | GRN | 0.122 | No | - | - | - |
| 58 | SAFSr_model | 0.08 | No | - | - | - |
| 59 | GAR-NOSF | 0.076 | No | - | - | - |
| 60 | SuppBERT | 0.056 | No | - | - | - |
| 61 | Baseline Model | 0.039 | No | HotpotQA: A Dataset for Diverse, Explainable Mul... | 2018-09-25 | Code |
| 62 | Entity-centric IR | 0.001 | No | - | - | - |
| 63 | tes | 0 | No | - | - | - |
| 64 | PromptRank-fewshot-2-demo | 0 | No | - | - | - |
| 65 | graph-recurrent-retriever+roberta-base w. S/R-pretraining | 0 | No | - | - | - |
| 66 | TPReasoner w/o BERT | 0 | No | - | - | - |
| 67 | MultiQA | 0 | No | - | - | - |
| 68 | DecompRC | 0 | No | - | - | Code |
| 69 | 0 | No | - | - | - | |
| 70 | Mistral multi hop with very large sources | 0 | No | - | - | - |