Last-word prediction on LAMBADA en
18.8Zero-shot AccMGPT-2
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MGPT-2#Params=124M, Patch size=-, zero-shot evaluation=true2026.04 | 18.8 | — | — | |
| GPT-2#Params=124M, Patch size=-, zero-shot evaluation=true2026.04 | 17.1 | — | — | |
| PIXAR stage2#Params=113M, Patch size=8×8, zero-shot evaluation=true2026.04 | 13.8 | 82.2 | — | |
| MIXAR stage2#Params=477M, Patch size=32×32, zero-shot evaluation=true2026.04 | 12.6 | 51 | 52.4 | |
| MIXAR stage2#Params=116M, Patch size=32×32, zero-shot evaluation=true2026.04 | 8 | 63.6 | 65.7 | |
| MIXAR stage1#Params=477M, Patch size=32×32, zero-shot evaluation=true2026.04 | 3.1 | 46 | 48.2 | |
| MIXAR stage1#Params=116M, Patch size=32×32, zero-shot evaluation=true2026.04 | 1.5 | 53.5 | 56.8 |