Question Answering on MMLU (AUROC, ECE)
0.87AUROCDISCOUQ-LLM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DISCOUQ-LLMID=M12026.03 | 0.87 | 0.049 | |
| DISCOUQ-LearnID=M32026.03 | 0.867 | 0.048 | |
| Embed Centr.ID=B52026.03 | 0.845 | 0.186 | |
| Vote CountID=B12026.03 | 0.841 | 0.114 | |
| DISCOUQ-EmbedID=M22026.03 | 0.84 | 0.056 | |
| Vote EntropyID=B22026.03 | 0.837 | 0.205 | |
| Self-Consist.ID=B42026.03 | 0.837 | 0.205 | |
| LLM Agg.ID=B62026.03 | 0.822 | 0.059 | |
| Verb. Conf.ID=B32026.03 | 0.72 | 0.143 |