# Caption-level correlation with human judgment on Flickr 8K tie-free (test)

Source: [Caption-level correlation with human judgment on Flickr 8K tie-free (test)](<https://www.sota2.com/research/sota/caption-level-correlation-with-human-judgment-on-flickr-8k-tie-free-test>)

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

Data updated: 2026-02-26T05:47:50.553Z

Evaluation of model performance on the test split of the Flickr 8K tie-free dataset, specifically measuring caption-level correlation against human judgments using metrics like Kendall's tau.

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 | Caption-level correlation with human judgment |
| Dataset (including any stated split) | Flickr 8K tie-free (test) |
| Metric | Kendall's Tau |
| Unit | SCALAR |
| 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

- [Kendall's Tau](<https://www.sota2.com/research/sota/caption-level-correlation-with-human-judgment-on-flickr-8k-tie-free-test.md?metric=Kendall%27s%20Tau>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Kendall's Tau | Result date | Report paper |
| --- | --- | --- | --- |
| SPARCS | 0.716 | Not recorded | 68fa6b991e209131cea2f2c0 |
| Inter-Human | 0.655 | Not recorded | 68fa6b991e209131cea2f2c0 |
| CIDEr | 0.65 | Not recorded | 68fa6b991e209131cea2f2c0 |
| SPICE | 0.649 | Not recorded | 68fa6b991e209131cea2f2c0 |
| METEOR | 0.611 | Not recorded | 68fa6b991e209131cea2f2c0 |
| Bleu-2 | 0.577 | Not recorded | 68fa6b991e209131cea2f2c0 |
| BERTScore | 0.53 | Not recorded | 68fa6b991e209131cea2f2c0 |
| Bleu-1 | 0.527 | Not recorded | 68fa6b991e209131cea2f2c0 |
| Rouge-L | 0.511 | Not recorded | 68fa6b991e209131cea2f2c0 |

### SPARCS

Kendall's Tau: 0.716. Result ID: 698e9868ffc3901de09ef9bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Implementation](<https://github.com/JoshuaFeinglass/SMURF>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### Inter-Human

Kendall's Tau: 0.655. Result ID: 698e9868ffc3901de09ef9be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a0f1e209131cea25e73.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### CIDEr

Kendall's Tau: 0.65. Result ID: 698e9868ffc3901de09ef9bb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa61e81e209131ce9f6a37.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### SPICE

Kendall's Tau: 0.649. Result ID: 698e9868ffc3901de09ef9bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62a41e209131ce9fa82e.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### METEOR

Kendall's Tau: 0.611. Result ID: 698e9868ffc3901de09ef9ba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### Bleu-2

Kendall's Tau: 0.577. Result ID: 698e9868ffc3901de09ef9b8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### BERTScore

Kendall's Tau: 0.53. Result ID: 698e9868ffc3901de09ef9b6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67171e209131cea140d5.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### Bleu-1

Kendall's Tau: 0.527. Result ID: 698e9868ffc3901de09ef9b7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

### Rouge-L

Kendall's Tau: 0.511. Result ID: 698e9868ffc3901de09ef9b9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6b991e209131cea2f2c0.md>)

Recorded evaluation configuration: {"Evaluation Level":"Caption-level"}. This record does not assert independent reproduction.

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