RAG Question Answering on SCORE domain-specific hazard response dataset (test)
61SpecificityGPT-4o
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-4oEvaluator=GPT-4o2026.02 | 61 | 89 | 60 | 69 | |
| GPT-4oEvaluator=Qwen3-8B2026.02 | 58 | 89 | 66 | 73 | |
| Gemini 2.5Evaluator=GPT-4o2026.02 | 55 | 87 | 87 | 68 | |
| Gemini 2.5Evaluator=Qwen3-8B2026.02 | 45 | 86 | 66 | 83 |