# Information Loss Question Answering Generation on INFOLOSSQA 1.0 (evaluation)

Source: [Information Loss Question Answering Generation on INFOLOSSQA 1.0 (evaluation)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation>)

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

Data updated: 2026-02-26T05:34:57.168Z

Evaluation of model performance on the INFOLOSSQA 1.0 dataset, focusing on Information Loss Question Answering Generation, measured by answerability metrics against source/target context and component accuracy (Evidence/Answer).

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 | Information Loss Question Answering Generation |
| Dataset (including any stated split) | INFOLOSSQA 1.0 (evaluation) |
| Metric | Answerable (Source Context Yes) |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| 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

- [Answerable (Source Context Yes)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Source%20Context%20Yes)>): Higher is better; unit: PERCENT.

- [Answerable (Source Context Multi)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Source%20Context%20Multi)>): Higher is better; unit: PERCENT.

- [Answerable (Source Context No)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Source%20Context%20No)>): Lower is better; unit: PERCENT.

- [Answerable (Target Context No)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Target%20Context%20No)>): Lower is better; unit: PERCENT.

- [Answerable (Target Context Vague)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Target%20Context%20Vague)>): Lower is better; unit: PERCENT.

- [Answerable (Target Context Yes)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Answerable%20(Target%20Context%20Yes)>): Higher is better; unit: PERCENT.

- [Acc (Evidence Correct)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Evidence%20Correct)>): Higher is better; unit: PERCENT.

- [Acc (Evidence Partial)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Evidence%20Partial)>): Lower is better; unit: PERCENT.

- [Acc (Evidence Wrong)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Evidence%20Wrong)>): Lower is better; unit: PERCENT.

- [Acc (Answer Correct)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Answer%20Correct)>): Higher is better; unit: PERCENT.

- [Acc (Answer Partial)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Answer%20Partial)>): Lower is better; unit: PERCENT.

- [Acc (Answer Wrong)](<https://www.sota2.com/research/sota/information-loss-question-answering-generation-on-infolossqa-1-0-evaluation.md?metric=Acc%20(Answer%20Wrong)>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Answerable (Source Context Yes) | Result date | Report paper |
| --- | --- | --- | --- |
| NLI Pipeline | 89.7 | Not recorded | 68fa73771e209131cea5cb44 |
| GPT-4 | 87.5 | Not recorded | 68fa73771e209131cea5cb44 |
| Llama2-70B | 83 | Not recorded | 68fa73771e209131cea5cb44 |
| Mistral-7B | 72.5 | Not recorded | 68fa73771e209131cea5cb44 |

### NLI Pipeline

Answerable (Source Context Yes): 89.7. Result ID: 6992e2563a6c31484912291e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73771e209131cea5cb44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73771e209131cea5cb44.md>)

[Implementation](<https://infolossqa.ikim.nrw/>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GPT-4

Answerable (Source Context Yes): 87.5. Result ID: 6992e2563a6c31484912291d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73771e209131cea5cb44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Llama2-70B

Answerable (Source Context Yes): 83. Result ID: 6992e2563a6c31484912291c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73771e209131cea5cb44.md>)

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

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

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Mistral-7B

Answerable (Source Context Yes): 72.5. Result ID: 6992e2563a6c31484912291b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73771e209131cea5cb44.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70d61e209131cea4d909.md>)

[Implementation](<https://github.com/mistralai/mistral-src>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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