Language Modeling on BabyLM STRICT 100M words (test)
82.8BLIMPELC-BERT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| ELC-BERTtraining_data=100M words, track=STRICT2023.11 | 82.8 | 78.3 | 47.2 | 74.3 | |
| BootBERTtraining_data=100M words, track=STRICT2023.11 | 82.2 | 78.5 | 27.7 | 70.2 | |
| Contextualizertraining_data=100M words, track=STRICT2023.11 | 79 | 72.9 | 58 | 73 | |
| Bad babiestraining_data=100M words, track=STRICT2023.11 | 77 | 67.2 | 23.4 | 63.4 | |
| MSLMtraining_data=100M words, track=STRICT2023.11 | 76.2 | 73.5 | 21.4 | 64.4 |