# Multi-hop Question Answering on MuSiQue 2022 Source: [Multi-hop Question Answering on MuSiQue 2022]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:43:22.403Z Evaluation of model performance on the MuSiQue 2022 dataset specifically for Multi-hop Question Answering, measured using metrics like F1, Tokens, and 1/T. 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 | Multi-hop Question Answering | | Dataset (including any stated split) | MuSiQue 2022 | | Metric | F1 Score | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 - [F1 Score](): Higher is better; unit: PERCENT. - [Tokens Count](): Higher is better; unit: SCALAR. - [Inverse Time (1/T)](): Higher is better; unit: TIME. ## Recorded results | Method | F1 Score | Result date | Report paper | | --- | --- | --- | --- | | LongLLMLingua | 51.2 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o Subsequence Recovery) | 50.7 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o restrict) | 49.2 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o Document Reordering) | 49.2 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o Question-aware Fine-grained) | 48.4 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/ GPT2-small) | 48.4 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o Dynamic Compression Ratio) | 48.2 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/ SBERT) | 47.3 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua r\_k | 46.3 | Not recorded | 68fa74701e209131cea61c6d | | Original Prompt | 45.8 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/ p(x\_doc\|x\_que, restrict)) | 44 | Not recorded | 68fa74701e209131cea61c6d | | LongLLMLingua (w/o Question-awareness) | 43.2 | Not recorded | 68fa74701e209131cea61c6d | | LLMLingua | 40.1 | Not recorded | 68fa74701e209131cea61c6d | | SBERT | 36.2 | Not recorded | 68fa74701e209131cea61c6d | | BM25 | 28.5 | Not recorded | 68fa74701e209131cea61c6d | | Selective-Context | 19.6 | Not recorded | 68fa74701e209131cea61c6d | ### LongLLMLingua F1 Score: 51.2. Result ID: 69935d283a6c314849156a6e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o Subsequence Recovery) F1 Score: 50.7. Result ID: 69935d283a6c314849156a75. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o restrict) F1 Score: 49.2. Result ID: 69935d283a6c314849156a72. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x","Module":"Question-aware Coarse-grained"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o Document Reordering) F1 Score: 49.2. Result ID: 69935d283a6c314849156a76. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o Question-aware Fine-grained) F1 Score: 48.4. Result ID: 69935d283a6c314849156a73. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### LongLLMLingua (w/ GPT2-small) F1 Score: 48.4. Result ID: 69935d283a6c314849156a77. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x","Model":"GPT2-small"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o Dynamic Compression Ratio) F1 Score: 48.2. Result ID: 69935d283a6c314849156a74. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### LongLLMLingua (w/ SBERT) F1 Score: 47.3. Result ID: 69935d283a6c314849156a70. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x","Module":"Question-aware Coarse-grained"}. This record does not assert independent reproduction. ### LongLLMLingua r\_k F1 Score: 46.3. Result ID: 69935d283a6c314849156a6b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### Original Prompt F1 Score: 45.8. Result ID: 69935d283a6c314849156a68. [Reporting paper]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo"}. This record does not assert independent reproduction. ### LongLLMLingua (w/ p(x\_doc\|x\_que, restrict)) F1 Score: 44. Result ID: 69935d283a6c314849156a71. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x","Module":"Question-aware Coarse-grained"}. This record does not assert independent reproduction. ### LongLLMLingua (w/o Question-awareness) F1 Score: 43.2. Result ID: 69935d283a6c314849156a6f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x","Module":"Question-aware Coarse-grained"}. This record does not assert independent reproduction. ### LLMLingua F1 Score: 40.1. Result ID: 69935d283a6c314849156a6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### SBERT F1 Score: 36.2. Result ID: 69935d283a6c314849156a6a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### BM25 F1 Score: 28.5. Result ID: 69935d283a6c314849156a69. [Reporting paper]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. ### Selective-Context F1 Score: 19.6. Result ID: 69935d283a6c314849156a6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Target LLM":"GPT-3.5-Turbo","Compression constraint":"2x"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.