RESULTS.md

Eval results:

Genia test set. First 600 subset. Claude 3.5 Sonnet. Simple prompting:

          Sample eval results with subset_size of 600

              precision    recall  f1-score   support

         DNA       0.41      0.55      0.47       387
         RNA       0.41      0.41      0.41        41
   cell_line       0.00      0.00      0.00         0
   cell_type       0.58      0.38      0.46       269
     protein       0.53      0.46      0.50       841

   micro avg       0.44      0.47      0.46      1538
   macro avg       0.39      0.36      0.37      1538
weighted avg       0.51      0.47      0.48      1538