Confidence Estimation on Global-MMLU Polish
77AUROCLast Answer
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Last AnswerProbe hidden state=last answer token, Training Language=French, Test Language=Polish, Zero-shot Transfer=true2026.05 | 77 | 47 | 0.32 | 0.36 | |
| Last QueryProbe hidden state=last query token, Training Language=French, Test Language=Polish, Zero-shot Transfer=true2026.05 | 69 | 39 | 0.21 | 0.2 |