# Unsupervised Machine Translation on English-Centric Directions Average (test)

Source: [Unsupervised Machine Translation on English-Centric Directions Average (test)](<https://www.sota2.com/research/sota/unsupervised-machine-translation-on-english-centric-directions-average-test>)

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

Data updated: 2026-02-26T06:06:34.452Z

Evaluation of Unsupervised Machine Translation performance on the English-Centric Directions Average test set, measured using Tokenized BLEU and Delta 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 | Unsupervised Machine Translation |
| Dataset (including any stated split) | English-Centric Directions Average (test) |
| Metric | Tokenized BLEU |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Tokenized BLEU](<https://www.sota2.com/research/sota/unsupervised-machine-translation-on-english-centric-directions-average-test.md?metric=Tokenized%20BLEU>): Higher is better; unit: SCALAR.

- [Delta](<https://www.sota2.com/research/sota/unsupervised-machine-translation-on-english-centric-directions-average-test.md?metric=Delta>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Tokenized BLEU | Result date | Report paper |
| --- | --- | --- | --- |
| mRASP2 | 18.55 | Not recorded | 68fa6a761e209131cea28501 |
| mRASP | 5.4 | Not recorded | 68fa6a761e209131cea28501 |
| m-Transformer | 4.42 | Not recorded | 68fa6a761e209131cea28501 |

### mRASP2

Tokenized BLEU: 18.55. Result ID: 698e3854ffc3901de09bc112.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a761e209131cea28501.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a761e209131cea28501.md>)

[Implementation](<https://github.com/PANXiao1994/mRASP2>)

Recorded evaluation configuration: {"training\_regime":"unsupervised"}. This record does not assert independent reproduction.

### mRASP

Tokenized BLEU: 5.4. Result ID: 698e3854ffc3901de09bc111.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a761e209131cea28501.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69161e209131cea201ee.md>)

[Implementation](<https://github.com/linzehui/mRASP>)

Recorded evaluation configuration: {"training\_regime":"unsupervised"}. This record does not assert independent reproduction.

### m-Transformer

Tokenized BLEU: 4.42. Result ID: 698e3854ffc3901de09bc110.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a761e209131cea28501.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa639f1e209131ce9ffc6d.md>)

Recorded evaluation configuration: {"training\_regime":"unsupervised"}. This record does not assert independent reproduction.

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