| Paper | Code | Accuracy | ModelName | ReleaseDate |
|---|---|---|---|---|
| OPT: Open Pre-trained Transformer Language Models | ✓ Link | 87 | OPT-175B | 2022-05-02 |
| Taylor-Calibrate: Principled Initialization for Hybrid Linear Attention Distillation | ✓ Link | 87.0 | Taylor-Calibrate qwen3_8b 100M_ga_s2_taylor | 2026-06-15 |
| Synthetic Pre-pretraining Survives Scale, but Not as a Grammatical Prior | ✓ Link | 85.0 | PPT (k-Shuffle Dyck), 3B, Marin, 100B PT tokens | 2026-09-30 |
| MultiHashFormer: Hash-based Generative Language Models | ✓ Link | 71.0 | MHF (H3B10K) | 2026-06-26 |
| Parcae: Scaling Laws For Stable Looped Language Models | ✓ Link | 68.666667 | Parcae-1.3B (T=8, 0-shot) | 2026-04-14 |