Question Answering on DROP 2048 tokens
30.2AccuracyLooped Hybrid (Full+GDN)
Evaluation Results
| Method | Links | |
|---|---|---|
| Looped Hybrid (Full+GDN)Model (1.5B)=1.5B, Looped=true2026.05 | 30.2 | |
| Looped Mamba-2Model (1.5B)=1.5B, Looped=true2026.05 | 29.7 | |
| Mamba-2Model (1.5B)=1.5B2026.05 | 28.5 | |
| Looped Hybrid (GDN+DSA)Model (1.5B)=1.5B, Looped=true2026.05 | 28.4 | |
| Looped TransformerModel (1.5B)=1.5B, Looped=true2026.05 | 28.1 | |
| TransformerModel (1.5B)=1.5B2026.05 | 26.4 | |
| Looped GDNModel (1.5B)=1.5B, Looped=true2026.05 | 25.9 | |
| GDNModel (1.5B)=1.5B2026.05 | 24.5 |