# Multi-modal Manipulation Detection and Grounding on DGM4 Wash. domain (test)

Source: [Multi-modal Manipulation Detection and Grounding on DGM4 Wash. domain (test)](<https://www.sota2.com/research/sota/multi-modal-manipulation-detection-and-grounding-on-dgm4-wash-domain-test>)

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

Data updated: 2026-03-04T16:44:53.903Z

Evaluation of Multi-modal Manipulation Detection and Grounding performance on the test split of the DGM4 Wash. domain dataset, reporting ACC, mAP, Ptok, and mIoU.

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-modal Manipulation Detection and Grounding |
| Dataset (including any stated split) | DGM4 Wash. domain (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/multi-modal-manipulation-detection-and-grounding-on-dgm4-wash-domain-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [mAP](<https://www.sota2.com/research/sota/multi-modal-manipulation-detection-and-grounding-on-dgm4-wash-domain-test.md?metric=mAP>): Higher is better; unit: PERCENT.

- [Ptok](<https://www.sota2.com/research/sota/multi-modal-manipulation-detection-and-grounding-on-dgm4-wash-domain-test.md?metric=Ptok>): Higher is better; unit: PERCENT.

- [mIoU](<https://www.sota2.com/research/sota/multi-modal-manipulation-detection-and-grounding-on-dgm4-wash-domain-test.md?metric=mIoU>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| REFORM | 72.84 | Not recorded | 69a66c552d603e6de3e3d0dd |
| AMD | 70.28 | Not recorded | 69a66c552d603e6de3e3d0dd |
| FKA-Owl | 67.05 | Not recorded | 69a66c552d603e6de3e3d0dd |
| HAMMER++ | 63.83 | Not recorded | 69a66c552d603e6de3e3d0dd |
| HAMMER | 63.54 | Not recorded | 69a66c552d603e6de3e3d0dd |
| LLaVa-v1.6-7B | 63.16 | Not recorded | 69a66c552d603e6de3e3d0dd |
| ViLT | 55.76 | Not recorded | 69a66c552d603e6de3e3d0dd |
| Qwen2.5-3B | 51.79 | Not recorded | 69a66c552d603e6de3e3d0dd |

### REFORM

Accuracy: 72.84. Result ID: 69a861786dbb7075b94ff640.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### AMD

Accuracy: 70.28. Result ID: 69a861786dbb7075b94ff63f.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b61e209131cea79f8e.md>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### FKA-Owl

Accuracy: 67.05. Result ID: 69a861786dbb7075b94ff63e.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa745e1e209131cea61338.md>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### HAMMER++

Accuracy: 63.83. Result ID: 69a861786dbb7075b94ff63d.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ab1e209131cea4c614.md>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### HAMMER

Accuracy: 63.54. Result ID: 69a861786dbb7075b94ff63c.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f091e209131cea43773.md>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### LLaVa-v1.6-7B

Accuracy: 63.16. Result ID: 69a861786dbb7075b94ff63a.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa717d1e209131cea51970.md>)

[Implementation](<https://llava-vl.github.io/>)

Recorded evaluation configuration: {"Category":"Fine-tuned LVLMs"}. This record does not assert independent reproduction.

### ViLT

Accuracy: 55.76. Result ID: 69a861786dbb7075b94ff63b.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a1e1e209131cea260db.md>)

[Implementation](<https://github.com/dandelin/vilt>)

Recorded evaluation configuration: {"Category":"Deepfake Detectors"}. This record does not assert independent reproduction.

### Qwen2.5-3B

Accuracy: 51.79. Result ID: 69a861786dbb7075b94ff639.

[Reporting paper](<https://www.sota2.com/research/paper/69a66c552d603e6de3e3d0dd.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773a1e209131cea71d3c.md>)

Recorded evaluation configuration: {"Category":"Fine-tuned LVLMs"}. This record does not assert independent reproduction.

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