# Multi-hop Question Answering on MuSiQue v1.0 (test)

Source: [Multi-hop Question Answering on MuSiQue v1.0 (test)](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-musique-v1-0-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:06:07.935Z

Evaluation of model latency (Task latency (s)) on the test split of the MuSiQue v1.0 dataset for Multi-hop Question Answering.

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 v1.0 (test) |
| Metric | Task latency (s) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 9 of 9 |
| 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

- [Task latency (s)](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-musique-v1-0-test.md?metric=Task%20latency%20(s)>): Lower is better; unit: TIME.

## Recorded results

| Method | Task latency (s) | Result date | Report paper |
| --- | --- | --- | --- |
| No Retrieval | 2.3 | Not recorded | 68fa790a1e209131cea7bd54 |
| VanillaRAG | 2.55 | Not recorded | 68fa790a1e209131cea7bd54 |
| L-RAG | 2.62 | Not recorded | 68fa790a1e209131cea7bd54 |
| VanillaRAG | 2.67 | Not recorded | 68fa790a1e209131cea7bd54 |
| L-RAG | 2.86 | Not recorded | 68fa790a1e209131cea7bd54 |
| HyDE | 4.81 | Not recorded | 68fa790a1e209131cea7bd54 |
| IRCoT | 14.57 | Not recorded | 68fa790a1e209131cea7bd54 |
| SelfAsk | 19.62 | Not recorded | 68fa790a1e209131cea7bd54 |
| IterRetGen | 21.18 | Not recorded | 68fa790a1e209131cea7bd54 |

### No Retrieval

Task latency (s): 2.3. Result ID: 6995bf4c3badfaffe6c18969.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B"}. This record does not assert independent reproduction.

### VanillaRAG

Task latency (s): 2.55. Result ID: 6995bf4c3badfaffe6c1896a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b91e209131cea23c26.md>)

[Implementation](<https://github.com/huggingface/transformers/blob/master/examples/rag/>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B","k":"4"}. This record does not assert independent reproduction.

### L-RAG

Task latency (s): 2.62. Result ID: 6995bf4c3badfaffe6c18970.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Implementation](<https://anonymous.4open.science/r/L-RAG-ADD5/>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B","k":"4"}. This record does not assert independent reproduction.

### VanillaRAG

Task latency (s): 2.67. Result ID: 6995bf4c3badfaffe6c1896b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69b91e209131cea23c26.md>)

[Implementation](<https://github.com/huggingface/transformers/blob/master/examples/rag/>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B","k":"8"}. This record does not assert independent reproduction.

### L-RAG

Task latency (s): 2.86. Result ID: 6995bf4c3badfaffe6c18971.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Implementation](<https://anonymous.4open.science/r/L-RAG-ADD5/>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B","k":"8"}. This record does not assert independent reproduction.

### HyDE

Task latency (s): 4.81. Result ID: 6995bf4c3badfaffe6c1896f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e3b1e209131cea3e835.md>)

[Implementation](<https://github.com/texttron/hyde>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B"}. This record does not assert independent reproduction.

### IRCoT

Task latency (s): 14.57. Result ID: 6995bf4c3badfaffe6c1896e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6fd81e209131cea4791d.md>)

[Implementation](<https://github.com/stonybrooknlp/ircot>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B"}. This record does not assert independent reproduction.

### SelfAsk

Task latency (s): 19.62. Result ID: 6995bf4c3badfaffe6c1896c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ef1e209131cea4e227.md>)

[Implementation](<https://github.com/ofirpress/self-ask>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B"}. This record does not assert independent reproduction.

### IterRetGen

Task latency (s): 21.18. Result ID: 6995bf4c3badfaffe6c1896d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa790a1e209131cea7bd54.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70f21e209131cea4e48a.md>)

Recorded evaluation configuration: {"Backbone":"Llama3.1-8B"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.