Confidence Estimation on Global-MMLU Russian (test)
75AUROCSeq. Likelihood
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Seq. LikelihoodBackbone=Qwen 3 8B, Evaluation Protocol=Zero-shot2026.05 | 75 | 47 | 54 | 62 | |
| Verbalized Conf.Backbone=Qwen 3 8B, Evaluation Protocol=Zero-shot2026.05 | 73 | 49 | 51 | 50 | |
| Cross-lingual ProbeBackbone=Qwen 3 8B, Evaluation Protocol=Zero-shot2026.05 | 72 | 43 | 39 | 42 | |
| Mass-Mean ProbeBackbone=Qwen 3 8B, Evaluation Protocol=Zero-shot2026.05 | 62 | 57 | 51 | 4 | |
| P(True)Backbone=Qwen 3 8B, Evaluation Protocol=Zero-shot2026.05 | 45 | 21 | 21 | 21 |