# Headline Generation on EXEQ-300k (test)

Source: [Headline Generation on EXEQ-300k (test)](<https://www.sota2.com/research/sota/headline-generation-on-exeq-300k-test>)

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

Data updated: 2026-02-26T05:37:17.719Z

Evaluation of headline generation performance on the test split of the EXEQ-300k dataset, assessed using standard sequence generation metrics including BLEU-4, ROUGE-L, and METEOR.

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 | Headline Generation |
| Dataset (including any stated split) | EXEQ-300k (test) |
| Metric | BLEU-4 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [BLEU-4](<https://www.sota2.com/research/sota/headline-generation-on-exeq-300k-test.md?metric=BLEU-4>): Higher is better; unit: PERCENT.

- [ROUGE-L](<https://www.sota2.com/research/sota/headline-generation-on-exeq-300k-test.md?metric=ROUGE-L>): Higher is better; unit: PERCENT.

- [METEOR](<https://www.sota2.com/research/sota/headline-generation-on-exeq-300k-test.md?metric=METEOR>): Higher is better; unit: PERCENT.

## Recorded results

| Method | BLEU-4 | Result date | Report paper |
| --- | --- | --- | --- |
| MathGPT | 56.5 | Not recorded | 68fa6e8d1e209131cea40850 |
| GPT-2 Wiki | 56.1 | Not recorded | 68fa6e8d1e209131cea40850 |
| GPT-2 | 55.3 | Not recorded | 68fa6e8d1e209131cea40850 |
| MathSum | 52 | Not recorded | 68fa6e8d1e209131cea40850 |
| MathBERT | 49.4 | Not recorded | 68fa6e8d1e209131cea40850 |

### MathGPT

BLEU-4: 56.5. Result ID: 699027fe3a6c314849fe805e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

[Implementation](<https://github.com/umass-ml4ed/mathGPT>)

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

### GPT-2 Wiki

BLEU-4: 56.1. Result ID: 699027fe3a6c314849fe805d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

Recorded evaluation configuration: {"math\_processing":"OPT pipeline","pre-training":"Wikipedia"}. This record does not assert independent reproduction.

### GPT-2

BLEU-4: 55.3. Result ID: 699027fe3a6c314849fe805c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

Recorded evaluation configuration: {"math\_processing":"OPT pipeline"}. This record does not assert independent reproduction.

### MathSum

BLEU-4: 52. Result ID: 699027fe3a6c314849fe805b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66cb1e209131cea12536.md>)

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

### MathBERT

BLEU-4: 49.4. Result ID: 699027fe3a6c314849fe805a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e8d1e209131cea40850.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69d21e209131cea246a8.md>)

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

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