# Negation Understanding on Neg-COCO MCQ

Source: [Negation Understanding on Neg-COCO MCQ](<https://www.sota2.com/research/sota/negation-understanding-on-neg-coco-mcq>)

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

Data updated: 2026-03-12T06:53:39.665Z

Evaluation of Negation Understanding performance on the Neg-COCO MCQ dataset, measured using Accuracy.

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 | Negation Understanding |
| Dataset (including any stated split) | Neg-COCO MCQ |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 14 of 14 |
| 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/negation-understanding-on-neg-coco-mcq.md?metric=Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DCSM | 48.6 | Not recorded | 68fa77651e209131cea73079 |
| CLIP | 41.5 | Not recorded | 68fa77651e209131cea73079 |
| CLIP | 39.2 | Not recorded | 68fa77651e209131cea73079 |
| DCSM | 38.8 | Not recorded | 68fa77651e209131cea73079 |
| CLIPGlasses | 35.9 | Not recorded | 699e82166017c4460ffacfcc |
| CLIPGlasses | 34.51 | Not recorded | 699e82166017c4460ffacfcc |
| NegCLIP | 31.4 | Not recorded | 68fa77651e209131cea73079 |
| CoN-CLIP | 30.61 | Not recorded | 699e82166017c4460ffacfcc |
| SigLIP | 26.3 | Not recorded | 68fa77651e209131cea73079 |
| CoN-CLIP | 25.7 | Not recorded | 699e82166017c4460ffacfcc |
| Random Chance | 25 | Not recorded | 68fa77651e209131cea73079 |
| CoCa | 21.6 | Not recorded | 68fa77651e209131cea73079 |
| BLIP | 20.5 | Not recorded | 68fa77651e209131cea73079 |
| NegCLIP | 10.2 | Not recorded | 699e82166017c4460ffacfcc |

### DCSM

Accuracy: 48.6. Result ID: 69b262e76dbb7075b95c6a93.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

Recorded evaluation configuration: {"Training Source":"coco","Base Model":"CLIP ViTB/16"}. This record does not assert independent reproduction.

### CLIP

Accuracy: 41.5. Result ID: 69b262e76dbb7075b95c6a8d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT B/16"}. This record does not assert independent reproduction.

### CLIP

Accuracy: 39.2. Result ID: 69b262e76dbb7075b95c6a8c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT B/32"}. This record does not assert independent reproduction.

### DCSM

Accuracy: 38.8. Result ID: 69b262e76dbb7075b95c6a92.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

Recorded evaluation configuration: {"Training Source":"synth","Base Model":"CLIP ViTB/16"}. This record does not assert independent reproduction.

### CLIPGlasses

Accuracy: 35.9. Result ID: 699ebe2f5698665ec48da47e.

[Reporting paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

Recorded evaluation configuration: {"Training Set":"Neg-COCO-R","Training Size (Images/Captions)":"5K/22K"}. This record does not assert independent reproduction.

### CLIPGlasses

Accuracy: 34.51. Result ID: 699ebe2f5698665ec48da47c.

[Reporting paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

Recorded evaluation configuration: {"Training Set":"CC-Neg","Training Size (Images/Captions)":"188K/376K"}. This record does not assert independent reproduction.

### NegCLIP

Accuracy: 31.4. Result ID: 69b262e76dbb7075b95c6a8e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ee31e209131cea42a7d.md>)

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

### CoN-CLIP

Accuracy: 30.61. Result ID: 699ebe2f5698665ec48da47d.

[Reporting paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

Recorded evaluation configuration: {"Training Set":"Neg-COCO-R","Training Size (Images/Captions)":"5K/22K"}. This record does not assert independent reproduction.

### SigLIP

Accuracy: 26.3. Result ID: 69b262e76dbb7075b95c6a91.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ae1e209131cea4c7a4.md>)

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

### CoN-CLIP

Accuracy: 25.7. Result ID: 699ebe2f5698665ec48da47b.

[Reporting paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

Recorded evaluation configuration: {"Training Set":"CC-Neg","Training Size (Images/Captions)":"188K/376K"}. This record does not assert independent reproduction.

### Random Chance

Accuracy: 25. Result ID: 69b262e76dbb7075b95c6a94.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

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

### CoCa

Accuracy: 21.6. Result ID: 69b262e76dbb7075b95c6a8f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cc71e209131cea35ecd.md>)

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

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

### BLIP

Accuracy: 20.5. Result ID: 69b262e76dbb7075b95c6a90.

[Reporting paper](<https://www.sota2.com/research/paper/68fa77651e209131cea73079.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bd81e209131cea30874.md>)

[Implementation](<https://github.com/salesforce/BLIP>)

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

### NegCLIP

Accuracy: 10.2. Result ID: 699ebe2f5698665ec48da47a.

[Reporting paper](<https://www.sota2.com/research/paper/699e82166017c4460ffacfcc.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ee31e209131cea42a7d.md>)

Recorded evaluation configuration: {"Training Set":"COCO-caption","Training Size (Images/Captions)":"82K/31K"}. This record does not assert independent reproduction.

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