Validation timing detection on M-TESC Multilingual
56.89Overall Macro-F1MEGUMI
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MEGUMI2026.06 | 56.89 | 41.44 | 48.7 | 44.78 | |
| XLM-RoBERTa2026.06 | 55.1 | 40.24 | 50 | 44.6 | |
| GPT 4o NanoEvaluation Protocol=Zero-shot2026.06 | 54.04 | 39.95 | 66.46 | 49.9 | |
| mBERT2026.06 | 53.29 | 38.29 | 41.35 | 39.76 | |
| GPT 4o NanoEvaluation Protocol=3-shot2026.06 | 50.14 | 39.01 | 80.93 | 52.65 | |
| Random Baseline2026.06 | 49.03 | 34.41 | 50.02 | 40.77 | |
| Llama 3.1 8bEvaluation Protocol=LoRA2026.06 | 47.99 | 33.83 | 58.44 | 42.86 | |
| GPT 4o NanoEvaluation Protocol=CoT2026.06 | 44.16 | 36.19 | 86.88 | 51.09 | |
| Llama 3.1 8bEvaluation Protocol=Zero-shot2026.06 | 40.76 | 36.31 | 89.58 | 51.68 | |
| Llama 3.1 8bEvaluation Protocol=3-shot2026.06 | 31.53 | 34.81 | 96.29 | 51.13 |