Ranking on NTCIR
74.2nDCGNLE (literal)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| NLE (literal)Explanation Type=literal, LLM=LLaMA2-13B-Chat2024.02 | 74.2 | 34 | |
| PLCalibration type=Platt Linear2024.02 | 73.6 | 36.4 | |
| NCCalibration type=No Calibration2024.02 | 73.5 | 33.7 | |
| PCCalibration type=Post-hoc Calibration2024.02 | 73.5 | 33.7 | |
| PRCalibration type=Platt Ranking2024.02 | 72.8 | 32.8 | |
| FC (BERT)Calibration type=Full Calibration, Backbone=BERT2024.02 | 72.7 | 28.5 | |
| NLE (conditional)Explanation Type=conditional, LLM=LLaMA2-13B-Chat2024.02 | 72 | 32.2 | |
| FC (monoBERT)Calibration type=Full Calibration, Backbone=monoBERT2024.02 | 69.6 | 26.8 |