+-------------+--------+----------+-------------+----------+----------------+
|             | Random | Semantic | Tokens Only | Base LLM | Fine-Tuned LLM |
+-------------+--------+----------+-------------+----------+----------------+
|   Accuracy  |  50%   |   63%    |     63%     |   60%    |      67%       |
|  Precision  |  44%   |   66%    |     58%     |    -     |       -        |
|    Recall   |  50%   |   31%    |     56%     |    -     |       -        |
|   F1 Score  |  47%   |   42%    |     57%     |    -     |       -        |
| Specificity |  50%   |   88%    |     68%     |    -     |       -        |
|     NPV     |  56%   |   62%    |     67%     |    -     |       -        |
+-------------+--------+----------+-------------+----------+----------------+