Hallucination Detection on HotpotQA (Accuracy)
80.3AccuracyVal Loss
Evaluation Results
| Method | Links | |
|---|---|---|
| Val LossModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 80.3 | |
| FEPoIDModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 80.3 | |
| FEPoIDModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 79.92 | |
| CurvatureModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 79.57 | |
| Val LossModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 79 | |
| CurvatureModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 77.79 | |
| SNRModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 77.79 | |
| SNRModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 76.94 | |
| RGNModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 74.93 | |
| RGNModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 74.85 | |
| IDModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 74.85 | |
| IDModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 73.09 | |
| RankMEModel=LlaMA-3.2-3B, Forward horizon (w)=7, Representation extraction (FST)=true2026.05 | 70.83 | |
| RankMEModel=LlaMA-3.2-1B, Forward horizon (w)=3, Representation extraction (FST)=true2026.05 | 70.09 |