# Multi-hop Question Answering on HotpotQA (EM, TFR, FFR, F1, IB)

Source: [Multi-hop Question Answering on HotpotQA (EM, TFR, FFR, F1, IB)](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib>)

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

Data updated: 2026-02-27T05:59:57.997Z

Evaluation of model performance on the HotpotQA dataset for multi-hop question answering, reporting Exact Match (EM), Total F1 Recall (TFR), Fact F1 Recall (FFR), F1 score, and Information Balance (IB).

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) | HOTPOTQA |
| Metric | EM |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [EM](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib.md?metric=EM>): Higher is better; unit: PERCENT.

- [TFR](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib.md?metric=TFR>): Higher is better; unit: PERCENT.

- [FFR](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib.md?metric=FFR>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

- [Information Balance (IB)](<https://www.sota2.com/research/sota/multi-hop-question-answering-on-hotpotqa-em-tfr-ffr-f1-ib.md?metric=Information%20Balance%20(IB)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | EM | Result date | Report paper |
| --- | --- | --- | --- |
| Information Bottleneck (IB) | 26.1 | Not recorded | 68fa7b9f1e209131cea89666 |
| LLAMA2 | 25.9 | Not recorded | 68fa7b9f1e209131cea89666 |
| LONGLLMLINGUA | 23.9 | Not recorded | 68fa7b9f1e209131cea89666 |
| RANKGPT | 23.5 | Not recorded | 68fa7b9f1e209131cea89666 |
| Top 1 document | 23.2 | Not recorded | 68fa7b9f1e209131cea89666 |
| LLAMA2-13B | 18.5 | Not recorded | 68fa7b9f1e209131cea89666 |
| Top 5 documents | 18.3 | Not recorded | 68fa7b9f1e209131cea89666 |

### Information Bottleneck (IB)

EM: 26.1. Result ID: 69a132c935dfcabb00e51083.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Implementation](<https://github.com/zhukun1020/NoiseFilter_IB>)

Recorded evaluation configuration: {"words":"13.2"}. This record does not assert independent reproduction.

### LLAMA2

EM: 25.9. Result ID: 69a132c935dfcabb00e51082.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"words":"27.7"}. This record does not assert independent reproduction.

### LONGLLMLINGUA

EM: 23.9. Result ID: 69a132c935dfcabb00e51081.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c6d.md>)

[Implementation](<https://aka.ms/LLMLingua>)

Recorded evaluation configuration: {"words":"137.7"}. This record does not assert independent reproduction.

### RANKGPT

EM: 23.5. Result ID: 69a132c935dfcabb00e51080.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76721e209131cea6d0b1.md>)

Recorded evaluation configuration: {"words":"102.9"}. This record does not assert independent reproduction.

### Top 1 document

EM: 23.2. Result ID: 69a132c935dfcabb00e5107e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

Recorded evaluation configuration: {"words":"102.9"}. This record does not assert independent reproduction.

### LLAMA2-13B

EM: 18.5. Result ID: 69a132c935dfcabb00e5107d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701f1e209131cea492fe.md>)

[Implementation](<https://github.com/facebookresearch/llama>)

Recorded evaluation configuration: {"words":"0"}. This record does not assert independent reproduction.

### Top 5 documents

EM: 18.3. Result ID: 69a132c935dfcabb00e5107f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b9f1e209131cea89666.md>)

Recorded evaluation configuration: {"words":"514.7"}. This record does not assert independent reproduction.

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