Eligibility Criteria Amendment Prediction on AMEND LLM (test)
0.697AUROCFinetuned BioBERT_CAMLM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Finetuned BioBERT_CAMLMModel=BioBERT, Pre-training Strategy=CAMLM, Evaluation Protocol=End-to-end fine-tuning2026.01 | 0.697 | 0.634 | 0.644 | |
| BioBERT_CAMLM + LRModel=BioBERT, Pre-training Strategy=CAMLM, Classifier=Logistic Regression, Evaluation Protocol=Frozen embedding2026.01 | 0.691 | 0.632 | 0.639 | |
| Finetuned BioBERTModel=BioBERT, Evaluation Protocol=End-to-end fine-tuning2026.01 | 0.687 | 0.621 | 0.619 | |
| BioBERT_CAMLM + RFModel=BioBERT, Pre-training Strategy=CAMLM, Classifier=Random Forest, Evaluation Protocol=Frozen embedding2026.01 | 0.686 | 0.628 | 0.633 | |
| BioBERT + LRModel=BioBERT, Classifier=Logistic Regression, Evaluation Protocol=Frozen embedding2026.01 | 0.681 | 0.61 | 0.633 | |
| BioBERT + RFModel=BioBERT, Classifier=Random Forest, Evaluation Protocol=Frozen embedding2026.01 | 0.676 | 0.603 | 0.617 |