Question Answering on ConflictQA (test)
68.9Correctness (%)ConflictRAG
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| ConflictRAGSeeds=32026.05 | 68.9 | 49.8 | 70.5 | |
| CoT DetectionTemperature=02026.05 | 63.1 | 46 | 43.5 | |
| NLI-FilterTemperature=02026.05 | 57.8 | 42.8 | 35.8 | |
| CRAGTemperature=02026.05 | 54.1 | 41.5 | 24.8 | |
| RAG+RerankTemperature=02026.05 | 52.3 | 40.1 | 23.4 | |
| Standard RAGTemperature=02026.05 | 49.2 | 37.8 | 22.1 | |
| Self-RAGTemperature=02026.05 | 47.8 | 36.2 | 22.4 | |
| RAPS-DA2026.06 | 42.7 | — | — | |
| OPCD2026.06 | 40 | — | — | |
| Mixed SFT2026.06 | 39.7 | — | — | |
| TIP2026.06 | 38.4 | — | — | |
| Knowledgeable-R12026.06 | 36.5 | — | — | |
| GRPO w/ RAG2026.06 | 36.3 | — | — | |
| RAG prompting2026.06 | 34.3 | — | — | |
| Astute-RAG2026.06 | 16.9 | — | — | |
| CK-PLUG2026.06 | 7.9 | — | — |