# Visual Hallucination Evaluation on AMBER Discriminative (Acc, Prec, Rec, F1)

Source: [Visual Hallucination Evaluation on AMBER Discriminative (Acc, Prec, Rec, F1)](<https://www.sota2.com/research/sota/visual-hallucination-evaluation-on-amber-discriminative-acc-prec-rec-f1>)

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

Data updated: 2026-06-24T07:27:19.640Z

Evaluation of visual hallucination detection performance on the AMBER Discriminative dataset, reporting Accuracy, Precision, Recall, and F1 scores.

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 | Visual Hallucination Evaluation |
| Dataset (including any stated split) | AMBER Discriminative |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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/visual-hallucination-evaluation-on-amber-discriminative-acc-prec-rec-f1.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [Precision](<https://www.sota2.com/research/sota/visual-hallucination-evaluation-on-amber-discriminative-acc-prec-rec-f1.md?metric=Precision>): Higher is better; unit: PERCENT.

- [Recall](<https://www.sota2.com/research/sota/visual-hallucination-evaluation-on-amber-discriminative-acc-prec-rec-f1.md?metric=Recall>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/visual-hallucination-evaluation-on-amber-discriminative-acc-prec-rec-f1.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| HeRA | 89 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 89 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 88.9 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen3-14B | 88.8 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 87.7 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen2.5-7B | 87.6 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen3-8B | 87.3 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen3-4B | 86.5 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 86.5 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen2.5-14B | 86.2 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 85.9 | Not recorded | 6a3b64b2206542251d7c97e6 |
| LLama3-8B | 85.9 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 85.9 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 85.7 | Not recorded | 6a3b64b2206542251d7c97e6 |
| HeRA | 85.6 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Vicuna-13B | 84.8 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Vicuna-7B | 84.2 | Not recorded | 6a3b64b2206542251d7c97e6 |
| Qwen2.5-3B | 83.6 | Not recorded | 6a3b64b2206542251d7c97e6 |

### HeRA

Accuracy: 89. Result ID: 6a3b86bbfbc373592c438f36.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 89. Result ID: 6a3b86bbfbc373592c438f3a.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 88.9. Result ID: 6a3b86bbfbc373592c438f3c.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen3-14B"}. This record does not assert independent reproduction.

### Qwen3-14B

Accuracy: 88.8. Result ID: 6a3b86bbfbc373592c438f3b.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-14B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 87.7. Result ID: 6a3b86bbfbc373592c438f34.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### Qwen2.5-7B

Accuracy: 87.6. Result ID: 6a3b86bbfbc373592c438f33.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-7B"}. This record does not assert independent reproduction.

### Qwen3-8B

Accuracy: 87.3. Result ID: 6a3b86bbfbc373592c438f35.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-8B"}. This record does not assert independent reproduction.

### Qwen3-4B

Accuracy: 86.5. Result ID: 6a3b86bbfbc373592c438f2d.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 86.5. Result ID: 6a3b86bbfbc373592c438f2e.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### Qwen2.5-14B

Accuracy: 86.2. Result ID: 6a3b86bbfbc373592c438f39.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-14B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 85.9. Result ID: 6a3b86bbfbc373592c438f30.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Vicuna-7B"}. This record does not assert independent reproduction.

### LLama3-8B

Accuracy: 85.9. Result ID: 6a3b86bbfbc373592c438f31.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d11e209131cea6a028.md>)

[Implementation](<https://github.com/meta-llama/llama-agentic-system>)

Recorded evaluation configuration: {"Backbone":"LLama3-8B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 85.9. Result ID: 6a3b86bbfbc373592c438f32.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"LLama3-8B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 85.7. Result ID: 6a3b86bbfbc373592c438f38.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Vicuna-13B"}. This record does not assert independent reproduction.

### HeRA

Accuracy: 85.6. Result ID: 6a3b86bbfbc373592c438f2c.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Implementation](<https://github.com/aimagelab/HeRA>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-3B"}. This record does not assert independent reproduction.

### Vicuna-13B

Accuracy: 84.8. Result ID: 6a3b86bbfbc373592c438f37.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

Recorded evaluation configuration: {"Backbone":"Vicuna-13B"}. This record does not assert independent reproduction.

### Vicuna-7B

Accuracy: 84.2. Result ID: 6a3b86bbfbc373592c438f2f.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

Recorded evaluation configuration: {"Backbone":"Vicuna-7B"}. This record does not assert independent reproduction.

### Qwen2.5-3B

Accuracy: 83.6. Result ID: 6a3b86bbfbc373592c438f2b.

[Reporting paper](<https://www.sota2.com/research/paper/6a3b64b2206542251d7c97e6.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa767e1e209131cea6d6a7.md>)

Recorded evaluation configuration: {"Backbone":"Qwen2.5-3B"}. This record does not assert independent reproduction.

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