Internal
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Internal Development Cohort patient-wise (test)
88.89Accuracy
7
Internal generalization probe (test)
0.033MAE
6
Internal
45.1Accuracy
6
Internal
78.5AUROC
6
Internal Dataset (10-fold cross-validation)
3.66MAE
6
Internal Uncertain Frozen (test)
1Macro AUC
5
Internal Slice-level (val)
91AUROC
5
Internal (test)
0.799AUC
5
Internal (test)
0.812RCA AUC
5
internal dataset (test)
2.45PESQ (WB)
5
Internal (test)
4.16MOS
4
Internal (test)
89.9HCM
4
Internal (test)
95.3Sensitivity (Spec=0.9) NH
4
Internal (hold-out test)
0.111RECIST Error
4
internal-8M
68Accuracy
4
Internal (test)
36.4Preference A (%)
3
Internal (test)
83Accuracy
3
Internal (test)
88Accuracy
3
Internal (test)
4.8Semantics-based Empathy
3
Internal (test)
4.53Instruction Following Score
3
Internal (held-out set)
95F1 Score
2
Internal cohort (test)
77.29Macro-AUC
1
Internal (val)
81.3AP (c)
1