Hallucination Regression on CSL-Daily
-0.079Pearson CorrelationPerplexity
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| PerplexityModel Type=Gloss-based (GB)2025.10 | -0.079 | -0.082 | 0.12 | |
| Token entropyModel Type=Gloss-based (GB)2025.10 | -0.158 | -0.143 | 0.195 | |
| ConfidenceModel Type=Gloss-based (GB)2025.10 | -0.253 | -0.269 | 0.294 | |
| GroundingModel Type=Gloss-based (GB)2025.10 | -0.439 | -0.426 | 0.465 | |
| METAModel Type=Gloss-based (GB)2025.10 | -0.518 | -0.515 | 0.543 | |
| PerplexityModel Type=Gloss-free (GF)2025.10 | -0.594 | -0.661 | 0.719 | |
| Token entropyModel Type=Gloss-free (GF)2025.10 | -0.67 | -0.698 | 0.748 | |
| GroundingModel Type=Gloss-free (GF)2025.10 | -0.682 | -0.68 | 0.722 | |
| ConfidenceModel Type=Gloss-free (GF)2025.10 | -0.685 | -0.657 | 0.714 | |
| METAModel Type=Gloss-free (GF)2025.10 | -0.736 | -0.705 | 0.755 |