Relation Extraction on GIDS (test)
88.3Micro-F1 (Positive Class)Qwen2.5-0.5B GenTune
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-0.5B GenTuneModel=Qwen2.5-0.5B, Shots=2s2026.06 | 88.3 | |
| Best SLMPrompting=schema-enumerated, Selection=strongest tuned SLM per benchmark2026.06 | 88.3 | |
| RoB-largeBackbone=RoBERTa, Parameters=355M, Inference mode=fine-tuned per benchmark2026.06 | 85.6 | |
| RoB-baseBackbone=RoBERTa, Parameters=125M, Inference mode=fine-tuned per benchmark2026.06 | 85 | |
| Llama-3.2-3B GenTuneModel=Llama-3.2-3B, Shots=2s2026.06 | 84.4 | |
| Llama-3.2-3B GenTuneModel=Llama-3.2-3B, Shots=0s2026.06 | 82 | |
| Claude Sonnet 4.6Shots=0-shot2026.06 | 79.9 | |
| ClaudeInference mode=zero-shot2026.06 | 79.9 | |
| GPT-5.4Shots=0-shot2026.06 | 79.1 | |
| GPT-5.4Inference mode=zero-shot2026.06 | 79.1 |