# Referring Segmentation on HALLUSEGBENCH Referring

Source: [Referring Segmentation on HALLUSEGBENCH Referring](<https://www.sota2.com/research/sota/referring-segmentation-on-hallusegbench-referring>)

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

Data updated: 2026-04-24T07:43:46.675Z

Evaluation of Referring Segmentation performance on the HALLUSEGBENCH Referring dataset, measured using CMS and Delta IoU metrics.

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 | Referring Segmentation |
| Dataset (including any stated split) | HALLUSEGBENCH Referring |
| Metric | CMS Factual |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| 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

- [CMS Factual](<https://www.sota2.com/research/sota/referring-segmentation-on-hallusegbench-referring.md?metric=CMS%20Factual>): Higher is better; unit: PERCENT.

- [CMS Counterfactual](<https://www.sota2.com/research/sota/referring-segmentation-on-hallusegbench-referring.md?metric=CMS%20Counterfactual>): Higher is better; unit: PERCENT.

- [Delta IoU Textual](<https://www.sota2.com/research/sota/referring-segmentation-on-hallusegbench-referring.md?metric=Delta%20IoU%20Textual>): Higher is better; unit: PERCENT.

- [Delta IoU Visual](<https://www.sota2.com/research/sota/referring-segmentation-on-hallusegbench-referring.md?metric=Delta%20IoU%20Visual>): Higher is better; unit: PERCENT.

## Recorded results

| Method | CMS Factual | Result date | Report paper |
| --- | --- | --- | --- |
| VisionReasoner | 81.91 | Not recorded | 68fa79ba1e209131cea7f22e |
| PixelLM-7B | 47.48 | Not recorded | 68fa79ba1e209131cea7f22e |
| Seg-Zero | 44.82 | Not recorded | 68fa79ba1e209131cea7f22e |
| PixelLM-13B | 43.06 | Not recorded | 68fa79ba1e209131cea7f22e |
| GLaMM-7B | 41.96 | Not recorded | 68fa79ba1e209131cea7f22e |
| LISA-13B | 31.94 | Not recorded | 68fa79ba1e209131cea7f22e |
| LISA-7B | 30.8 | Not recorded | 68fa79ba1e209131cea7f22e |
| SESAME-7B | 19.83 | Not recorded | 68fa79ba1e209131cea7f22e |
| RobustSeg | 10.62 | Not recorded | 68fa79ba1e209131cea7f22e |

### VisionReasoner

CMS Factual: 81.91. Result ID: 69eb1f2847f34ce924d30d1d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7ba21e209131cea897a6.md>)

Recorded evaluation configuration: {"Hallucination Mitigation":"true"}. This record does not assert independent reproduction.

### PixelLM-7B

CMS Factual: 47.48. Result ID: 69eb1f2847f34ce924d30d18.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74241e209131cea5ff3b.md>)

[Implementation](<https://pixellm.github.io/>)

Recorded evaluation configuration: {"Hallucination Mitigation":"false"}. This record does not assert independent reproduction.

### Seg-Zero

CMS Factual: 44.82. Result ID: 69eb1f2847f34ce924d30d1c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79b81e209131cea7f19c.md>)

Recorded evaluation configuration: {"Hallucination Mitigation":"true"}. This record does not assert independent reproduction.

### PixelLM-13B

CMS Factual: 43.06. Result ID: 69eb1f2847f34ce924d30d1b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74241e209131cea5ff3b.md>)

[Implementation](<https://pixellm.github.io/>)

Recorded evaluation configuration: {"Hallucination Mitigation":"false"}. This record does not assert independent reproduction.

### GLaMM-7B

CMS Factual: 41.96. Result ID: 69eb1f2847f34ce924d30d19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73651e209131cea5c2d8.md>)

[Implementation](<https://mbzuai-oryx.github.io/groundingLMM>)

Recorded evaluation configuration: {"Hallucination Mitigation":"false"}. This record does not assert independent reproduction.

### LISA-13B

CMS Factual: 31.94. Result ID: 69eb1f2847f34ce924d30d1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Hallucination Mitigation":"false"}. This record does not assert independent reproduction.

### LISA-7B

CMS Factual: 30.8. Result ID: 69eb1f2847f34ce924d30d17.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa730a1e209131cea5a0b2.md>)

[Implementation](<https://www.sota2.com/github.com/dvlab-research/LISA>)

Recorded evaluation configuration: {"Hallucination Mitigation":"false"}. This record does not assert independent reproduction.

### SESAME-7B

CMS Factual: 19.83. Result ID: 69eb1f2847f34ce924d30d1e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71701e209131cea513b6.md>)

Recorded evaluation configuration: {"Hallucination Mitigation":"true"}. This record does not assert independent reproduction.

### RobustSeg

CMS Factual: 10.62. Result ID: 69eb1f2847f34ce924d30d1f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79ba1e209131cea7f22e.md>)

Recorded evaluation configuration: {"Hallucination Mitigation":"true"}. This record does not assert independent reproduction.

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