| Paper | Code | ROUGE-2 | BLEU score | Parameters | ModelName | ReleaseDate |
|---|---|---|---|---|---|---|
| The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics | 23.2 | 568 M | PEGASUS | 2021-02-02 | ||
| PaLM: Scaling Language Modeling with Pathways | ✓ Link | 21.2 | 540 B | PaLM (finetuning)-540B | 2022-04-05 | |
| PaLM: Scaling Language Modeling with Pathways | ✓ Link | 21.0 | T5-XXL | 2022-04-05 | ||
| PaLM: Scaling Language Modeling with Pathways | ✓ Link | 18.5 | 62 B | PaLM (finetuning)-62B | 2022-04-05 | |
| ByT5: Towards a token-free future with pre-trained byte-to-byte models | ✓ Link | 15.3 | ByT5 | 2021-05-28 | ||
| ByT5: Towards a token-free future with pre-trained byte-to-byte models | ✓ Link | 14.3 | mT5 | 2021-05-28 |