Validation timing detection on M-EDESConv Multilingual
70.82Macro-F1 (Overall)Human
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Human2026.06 | 70.82 | 66.2 | 88.67 | 75.78 | |
| MEGUMI2026.06 | 63.71 | 51.07 | 66.11 | 57.62 | |
| MEGUMI2026.06 | 61.15 | 64.2 | 52 | 57.46 | |
| mBERT2026.06 | 59.1 | 45.16 | 74.15 | 56.14 | |
| XLM-RoBERTa2026.06 | 59.03 | 45.19 | 76.96 | 56.94 | |
| XLM-RoBERTa2026.06 | 58.85 | 60.23 | 53 | 56.38 | |
| mBERT2026.06 | 58.8 | 60.47 | 52 | 55.91 | |
| GPT 4o NanoEvaluation Protocol=Zero-shot2026.06 | 51.76 | 52.14 | 46.8 | 49.31 | |
| GPT 4o NanoEvaluation Protocol=Zero-shot2026.06 | 51.68 | 41.43 | 79.25 | 54.41 | |
| GPT 4o NanoEvaluation Protocol=3-shot2026.06 | 49.5 | 53.01 | 88 | 66.17 | |
| Random Baseline2026.06 | 49.23 | 36.45 | 50.35 | 42.3 | |
| Random Baseline2026.06 | 48.68 | 48.71 | 47.6 | 48.13 | |
| GPT 4o NanoEvaluation Protocol=3-shot2026.06 | 46.75 | 40.02 | 87.04 | 54.99 | |
| GPT 4o NanoEvaluation Protocol=CoT2026.06 | 46.54 | 52.89 | 95.4 | 68.05 | |
| GPT 4o NanoEvaluation Protocol=CoT2026.06 | 42.19 | 39.16 | 92.88 | 55.09 | |
| Llama 3.1 8bEvaluation Protocol=LoRA2026.06 | 42.02 | 51.85 | 98 | 67.82 | |
| Llama 3.1 8bEvaluation Protocol=3-shot2026.06 | 39.12 | 51.04 | 98 | 67.12 | |
| Llama 3.1 8bEvaluation Protocol=Zero-shot2026.06 | 37.18 | 37.72 | 93.75 | 53.79 | |
| Llama 3.1 8bEvaluation Protocol=Zero-shot2026.06 | 36.54 | 50.16 | 96.8 | 66.07 | |
| Llama 3.1 8bEvaluation Protocol=3-shot2026.06 | 35.69 | 37.49 | 94.84 | 53.73 | |
| Llama 3.1 8bEvaluation Protocol=LoRA2026.06 | 34.42 | 36.4 | 90.81 | 51.97 |