+-------------+--------+----------+-------------+
|             | Random | Semantic | Tokens Only |
+-------------+--------+----------+-------------+
|   Accuracy  | 50.0%  |  80.2%   |    73.3%    |
|   Balanced  | 50.0%  |  67.3%   |    70.9%    |
|  Precision  | 25.7%  |  69.6%   |    48.4%    |
|    Recall   | 50.0%  |  40.7%   |    66.1%    |
|   F1 Score  | 33.9%  |  51.3%   |    55.9%    |
| Specificity | 50.0%  |  93.9%   |    75.7%    |
|     NPV     | 74.3%  |  82.1%   |    86.6%    |
+-------------+--------+----------+-------------+