Selective Prediction on WMT de 19
46.5PRRMeanTokenEntropy
Evaluation results
| Method | Links | |
|---|---|---|
| MeanTokenEntropy2026.04 | 46.5 | |
| Perplexity2026.04 | 46.1 | |
| RAUQ2026.04 | 45.8 | |
| SP×MeanEntropyGenerator=GPT, Optimization=Evolved specifically for selective prediction2026.04 | 45.4 | |
| Evo-Gpt-5Generator=GPT2026.04 | 45.4 | |
| Evo-Gpt-1Generator=GPT2026.04 | 45.4 | |
| Evo-Anth-30Generator=Claude2026.04 | 45.1 | |
| Evo-Anth-8Generator=Claude2026.04 | 45 | |
| Evo-Anth-1Generator=Claude2026.04 | 44.9 | |
| SP×ExpectedRankGenerator=GPT, Optimization=Evolved specifically for selective prediction2026.04 | 44.7 | |
| Evo-Gpt-9Generator=GPT2026.04 | 44 | |
| Evo-Gpt-7Generator=GPT2026.04 | 44 | |
| SequenceProbability2026.04 | 44 | |
| Evo-Anth-22Generator=Claude2026.04 | 44 | |
| Evo-Anth-15Generator=Claude2026.04 | 43.8 | |
| Evo-Gpt-13Generator=GPT2026.04 | 43.7 | |
| Evo-Gpt-11Generator=GPT2026.04 | 42.9 | |
| Evo-Gpt-15Generator=GPT2026.04 | 42.6 | |
| Evo-Gpt-3Generator=GPT2026.04 | 41.9 | |
| AttentionScore2026.04 | 40.7 |