| Benchmark | Model | Rank | Results |
|---|---|---|---|
| language-modelling-on-enwiki8 | Sandwich Transformer (adaptive span) | #6 | Bit per Character (BPC): 0.968Number of params: 209M |
| language-modelling-on-wikitext-103 | Sandwich Transformer | #28 | Test perplexity: 17.96Number of params: 247M |