Validation timing detection on M-EDESConv English
62.22Overall Macro-F1XLM-RoBERTa
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| XLM-RoBERTa2026.06 | 62.22 | 47.24 | 70.38 | 56.57 | |
| MEGUMI2026.06 | 61.67 | 48.4 | 60.99 | 53.97 | |
| BERT2026.06 | 59.74 | 45.16 | 74.15 | 56.19 | |
| ModernBERT2026.06 | 59.1 | 44.94 | 75.02 | 56.27 | |
| GPT 4o NanoEvaluation Protocol=3-shot2026.06 | 56.69 | 43.19 | 68.49 | 52.98 | |
| GPT 4o NanoEvaluation Protocol=Zero-shot2026.06 | 55.36 | 42.53 | 74.43 | 54.13 | |
| Llama 3.1 8bEvaluation Protocol=LoRA2026.06 | 50.18 | 33.33 | 31.17 | 32.21 | |
| Random Baseline2026.06 | 49.4 | 35.83 | 50.59 | 41.95 | |
| GPT 4o NanoEvaluation Protocol=CoT2026.06 | 43.84 | 38.88 | 92.39 | 54.73 | |
| Llama 3.1 8bEvaluation Protocol=Zero-shot2026.06 | 40.9 | 37.55 | 90.54 | 53.08 | |
| Llama 3.1 8bEvaluation Protocol=3-shot2026.06 | 33.75 | 36.45 | 96.47 | 52.91 |