| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| WikiText2 | Perplexity2.73 | 4,085 | 29d ago | ||
| WikiText-2 | Calibration-free SLEB | Perplexity (PPL)1.08 | 2,862 | 26d ago | |
| WikiText-2 (test) | PPL1.2 | 2,416 | 16d ago | ||
| C4 | Wanda | Perplexity1 | 1,688 | 2mo ago | |
| C4 | Llama2-7B | Perplexity4.77 | 1,565 | 2mo ago | |
| PTB | Perplexity4.345 | 1,234 | 2mo ago | ||
| C4 (val) | PPL1.7 | 908 | 22d ago | ||
| WikiText-103 (test) | RETRO | Perplexity2.22 | 773 | 26d ago | |
| WikiText | Llama 3.1-70B | PPL0.2838 | 740 | 2mo ago | |
| PTB (test) | Perplexity8.159 | 543 | 1mo ago | ||
| C4 (test) | Perplexity4.97 | 488 | 1mo ago | ||
| C4 | Perplexity5.52 | 482 | 18d ago | ||
| Penn Treebank (test) | GL-LWGC-AWD-MOS-LSTM + dynamic evaluation | Perplexity46.34 | 450 | 18d ago | |
| Wikitext2 | Perplexity2.58 | 447 | 1mo ago | ||
| WikiText2 (val) | TCN-SEQ-J | Perplexity (PPL)3.03 | 436 | 1mo ago | |
| LAMBADA | PaLM-2 L | Accuracy86.9 | 412 | 1mo ago | |
| WikiText2 v1 (test) | Perplexity1.7 | 396 | 1mo ago | ||
| Wiki2 | PPL4.01 | 382 | 1mo ago | ||
| WikiText | Word Perplexity2.9 | 331 | 1mo ago | ||
| Wiki | Wanda | Perplexity (PPL)2 | 298 | 3mo ago | |
| WikiText-103 (val) | PPL1.01 | 290 | 1mo ago | ||
| FineWeb (val) | Muon (Sync) | Validation Loss1.906 | 259 | 25d ago | |
| LAMBADA | Perplexity3.1 | 254 | 1mo ago | ||
| PG-19 | CREAM | Perplexity5 | 244 | 1mo ago | |
| WikiText-103 | ESPACE | PPL4.59 | 216 | 2mo ago |