Classification on ISARC LLM Encoding Study Dataset 2025 (42 folds cross-validation)
0.0066P-Valuetext-embedding-3-large
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| text-embedding-3-largeEncoding Format=compacted, Baseline Encoding=One-hot, Statistical Test=Wilcoxon signed-rank test2026.02 | 0.0066 | 2.7179 | — | |
| text-embedding-3-largeEncoding Format=original, Baseline Encoding=One-hot, Statistical Test=Wilcoxon signed-rank test2026.02 | 0.1639 | 1.3922 | — | |
| llama-3Encoding Format=original, Baseline Encoding=One-hot, Statistical Test=Wilcoxon signed-rank test2026.02 | 0.2193 | 1.2284 | — | |
| text-embedding-3-smallEncoding Format=compacted, Baseline Encoding=One-hot, Statistical Test=Wilcoxon signed-rank test2026.02 | 0.2853 | 1.0685 | — | |
| llama-3Encoding Format=compacted, Baseline Encoding=One-hot, Statistical Test=Wilcoxon signed-rank test2026.02 | 0.7141 | 0.3663 | — | |
| text-embedding-3-smallEncoding Format=original, Baseline Encoding=One-hot, Statistical Test=Paired t-test2026.02 | 0.8731 | — | -0.1607 |