# Region-level faithfulness analysis on DeFacto 1.5K (test)

Source: [Region-level faithfulness analysis on DeFacto 1.5K (test)](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test>)

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

Data updated: 2026-05-21T10:17:11.108Z

Evaluation of model faithfulness focusing on region-level analysis using the test split of the DeFacto 1.5K dataset, measured by metrics such as Mislocalized & Wrong, Spurious Correct, and Abstention rates.

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 | Region-level faithfulness analysis |
| Dataset (including any stated split) | DeFacto 1.5K (test) |
| Metric | Mislocalized & Wrong Rate |
| Unit | PERCENT |
| Direction | Lower is better |
| Results loaded | 3 of 3 |
| 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

- [Mislocalized & Wrong Rate](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test.md?metric=Mislocalized%20%26%20Wrong%20Rate>): Lower is better; unit: PERCENT.

- [Spurious Correct Rate](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test.md?metric=Spurious%20Correct%20Rate>): Lower is better; unit: PERCENT.

- [Faithful but Wrong Rate](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test.md?metric=Faithful%20but%20Wrong%20Rate>): Lower is better; unit: PERCENT.

- [Masked Evidence Abstention Rate](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test.md?metric=Masked%20Evidence%20Abstention%20Rate>): Lower is better; unit: PERCENT.

- [Image Replacement Abstention Rate](<https://www.sota2.com/research/sota/region-level-faithfulness-analysis-on-defacto-1-5k-test.md?metric=Image%20Replacement%20Abstention%20Rate>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Mislocalized & Wrong Rate | Result date | Report paper |
| --- | --- | --- | --- |
| DeFacto | 21.5 | Not recorded | 68fa7b991e209131cea8964c |
| DeepEyes | 23.5 | Not recorded | 68fa7b991e209131cea8964c |
| GRIT | 43.5 | Not recorded | 68fa7b991e209131cea8964c |

### DeFacto

Mislocalized & Wrong Rate: 21.5. Result ID: 6a0edb91fbc373592c06afac.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b991e209131cea8964c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b991e209131cea8964c.md>)

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

### DeepEyes

Mislocalized & Wrong Rate: 23.5. Result ID: 6a0edb91fbc373592c06afaa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b991e209131cea8964c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c71e209131cea7a75e.md>)

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

### GRIT

Mislocalized & Wrong Rate: 43.5. Result ID: 6a0edb91fbc373592c06afab.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b991e209131cea8964c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78bd1e209131cea7a2a3.md>)

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.