Hallucination Evaluation on ModeVent Inliers In-Domain interference 1.0
2.4Attribute Hijacking FidelityMODE-RAG
Evaluation results
| Method | Links | |||||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| MODE-RAG2026.06 | 2.4 | 1.26 | 3.66 | 2.64 | 1.39 | 4.03 | 3.14 | 1.71 | 4.86 | 3.4 | 2.83 | 6.23 | 3.8 | 2.02 | 5.82 | 3.31 | 1.94 | 5.25 | 2.89 | 1.9 | 4.79 | 3.07 | 1.84 | 4.91 | 0.7 | -0.1 | 0.6 | |
| MODE-RAG2026.06 | 2.39 | 1.44 | 3.82 | 2.71 | 1.81 | 4.52 | 3.34 | 1.88 | 5.22 | 3.67 | 3.16 | 6.83 | 3.76 | 2.4 | 6.16 | 3.67 | 2.57 | 6.24 | 3.11 | 2.34 | 5.45 | 3.23 | 2.22 | 5.45 | 0.89 | 0.16 | 1.04 | |
| Baseline2026.06 | 1.66 | 1.08 | 2.74 | 1.84 | 1.86 | 3.7 | 2.2 | 1.72 | 3.93 | 2.62 | 2.79 | 5.41 | 2.73 | 2.37 | 5.1 | 2.8 | 2.47 | 5.27 | 2.59 | 2.21 | 4.8 | 2.34 | 2.06 | 4.4 | — | — | — | |
| Baseline2026.06 | 1.45 | 0.85 | 2.3 | 1.82 | 1.78 | 3.6 | 2.32 | 1.61 | 3.92 | 2.83 | 2.98 | 5.82 | 2.94 | 2.32 | 5.26 | 2.86 | 2.19 | 5.05 | 2.56 | 2.13 | 4.69 | 2.37 | 1.94 | 4.31 | — | — | — |