Research Article
Self-Information Loss Compensation Learning for Machine-Generated Text Detection
Table 4
Comparison of results with two initialization parameter methods.
| Sequence length | 64 | 128 | 256 | 512 | Our model | F1 score | ACC | F1 score | ACC | F1 score | ACC | F1 score | ACC |
| Random initialization parameters | 0.8147 | 0.8283 | 0.8633 | 0.8683 | 0.8640 | 0.8683 | 0.8659 | 0.8708 | Pretrained parameters | 0.8731 | 0.8792 | 0.9072 | 0.9103 | 0.9003 | 0.9050 | 0.9131 | 0.9158 |
|
|