Language Modeling on WikiText-2 1M
1.2745Lossdual_film
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| dual_filmStructure variant=dual_film2026.06 | 1.2745 | 3.5768 | 4.96 | |
| single_localStructure variant=single_local2026.06 | 1.3067 | 3.6938 | 1.48 | |
| dual_concatStructure variant=dual_concat2026.06 | 1.3097 | 3.7051 | 2.39 | |
| single_globalStructure variant=single_global2026.06 | 1.3169 | 3.7319 | 1.25 |