Machine Translation Evaluation on Aligned 300-word human-evaluation chunks
8.657HT Mean ScoreAgentic (P3) MT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Agentic (P3) MTMetric=MetricX-QE, Chunks=3852026.06 | 8.657 | 8.347 | -0.309 | 0.128 | |
| Agentic (P3) MTMetric=MetricX-QE (paragraph mean), Chunks=3852026.06 | 5.671 | 5.184 | -0.486 | 0.103 | |
| Agentic (P3) MTMetric=LiTransProQA, Chunks=3862026.06 | 0.92 | 0.996 | -0.076 | 0.017 | |
| Agentic (P3) MTMetric=COMETKiwi (paragraph mean), Chunks=3772026.06 | 0.722 | 0.76 | -0.038 | 0.004 |