Relevance Classification on Relevance classification (test)
88.9AccuracyLLaMA-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLaMA-8BModel source=ours2026.06 | 88.9 | 200 | |
| Gemma-2BModel source=ours2026.06 | 88.2 | 50 | |
| Cross-EncoderModel source=ours2026.06 | 87.5 | 5 | |
| Claude-3Model source=off-the-shelf2026.06 | 70.1 | 450 | |
| GPT-4Model source=off-the-shelf2026.06 | 68.2 | 500 |