Selective Prediction on WMT fr 14
39.1Prediction Ranking RateMeanTokenEntropy
Evaluation results
| Method | Links | |
|---|---|---|
| MeanTokenEntropy2026.04 | 39.1 | |
| Perplexity2026.04 | 39 | |
| RAUQ2026.04 | 38.8 | |
| Evo-Gpt-1Generator=GPT2026.04 | 38.5 | |
| Evo-Gpt-5Generator=GPT2026.04 | 38.4 | |
| Evo-Anth-8Generator=Claude2026.04 | 38.1 | |
| SP×MeanEntropyGenerator=GPT, Optimization=Evolved specifically for selective prediction2026.04 | 38 | |
| Evo-Anth-1Generator=Claude2026.04 | 37.9 | |
| Evo-Anth-30Generator=Claude2026.04 | 37.8 | |
| SP×ExpectedRankGenerator=GPT, Optimization=Evolved specifically for selective prediction2026.04 | 37.6 | |
| Evo-Gpt-13Generator=GPT2026.04 | 37.2 | |
| Evo-Anth-22Generator=Claude2026.04 | 37.1 | |
| Evo-Gpt-9Generator=GPT2026.04 | 37 | |
| Evo-Gpt-7Generator=GPT2026.04 | 37 | |
| SequenceProbability2026.04 | 37 | |
| Evo-Anth-15Generator=Claude2026.04 | 36.9 | |
| Evo-Gpt-11Generator=GPT2026.04 | 36.7 | |
| Evo-Gpt-15Generator=GPT2026.04 | 36.4 | |
| Evo-Gpt-3Generator=GPT2026.04 | 35.7 | |
| AttentionScore2026.04 | 35.1 |