Generation Correctness Prediction on TruthfulQA (test)
62.69AURCDeg
Evaluation Results
| Method | Links | |
|---|---|---|
| DegBackbone=Llama3.1-Instruct8B2026.05 | 62.69 | |
| S.D.Backbone=Llama3.1-Instruct8B2026.05 | 62.38 | |
| P(True)Backbone=Llama3.1-Instruct8B2026.05 | 61.84 | |
| Self-ConBackbone=Llama3.1-Instruct8B2026.05 | 61.37 | |
| S.E.Backbone=Llama3.1-Instruct8B2026.05 | 60.76 | |
| INSIDEBackbone=Llama3.1-Instruct8B2026.05 | 60.43 | |
| G-NLLBackbone=Llama3.1-Instruct8B2026.05 | 59.08 | |
| P(True)Backbone=Mistral-Nemo-Instruct12B2026.05 | 58.22 | |
| P(True)Backbone=Qwen3-Instruct4B2026.05 | 57.24 | |
| ExGradBackbone=Llama3.1-Instruct8B2026.05 | 57.15 | |
| ParaGradBackbone=Llama3.1-Instruct8B2026.05 | 56.15 | |
| LN-PEBackbone=Llama3.1-Instruct8B2026.05 | 55.78 | |
| SARBackbone=Llama3.1-Instruct8B2026.05 | 54.35 | |
| M.I.Backbone=Llama3.1-Instruct8B2026.05 | 54.03 | |
| HybridGradBackbone=Llama3.1-Instruct8B2026.05 | 52.35 | |
| S.D.Backbone=Qwen3-Instruct4B2026.05 | 52.08 | |
| SemGradBackbone=Llama3.1-Instruct8B2026.05 | 51.99 | |
| INSIDEBackbone=Mistral-Nemo-Instruct12B2026.05 | 51.97 | |
| S.D.Backbone=Mistral-Nemo-Instruct12B2026.05 | 51.65 | |
| LN-PEBackbone=Mistral-Nemo-Instruct12B2026.05 | 50.97 | |
| G-NLLBackbone=Mistral-Nemo-Instruct12B2026.05 | 49.67 | |
| DegBackbone=Mistral-Nemo-Instruct12B2026.05 | 49.52 | |
| S.E.Backbone=Mistral-Nemo-Instruct12B2026.05 | 49.39 | |
| ExGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 48.96 | |
| Self-ConBackbone=Mistral-Nemo-Instruct12B2026.05 | 48.56 | |
| DegBackbone=Qwen3-Instruct4B2026.05 | 48.17 | |
| LN-PEBackbone=Qwen3-Instruct4B2026.05 | 47.74 | |
| SARBackbone=Mistral-Nemo-Instruct12B2026.05 | 47.49 | |
| M.I.Backbone=Qwen3-Instruct4B2026.05 | 46.56 | |
| M.I.Backbone=Mistral-Nemo-Instruct12B2026.05 | 46.2 | |
| ParaGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 46.16 | |
| S.E.Backbone=Qwen3-Instruct4B2026.05 | 45.97 | |
| G-NLLBackbone=Qwen3-Instruct4B2026.05 | 45.93 | |
| Self-ConBackbone=Qwen3-Instruct4B2026.05 | 44.87 | |
| SemGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 44.86 | |
| INSIDEBackbone=Qwen3-Instruct4B2026.05 | 44.61 | |
| HybridGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 44.54 | |
| ExGradBackbone=Qwen3-Instruct4B2026.05 | 44.45 | |
| ParaGradBackbone=Qwen3-Instruct4B2026.05 | 43.42 | |
| SARBackbone=Qwen3-Instruct4B2026.05 | 42.7 | |
| SemGradBackbone=Qwen3-Instruct4B2026.05 | 41.74 | |
| HybridGradBackbone=Qwen3-Instruct4B2026.05 | 41.58 |