Generation correctness prediction on SciQ (test)
35.01AURCS.E.
Evaluation Results
| Method | Links | |
|---|---|---|
| S.E.Backbone=Qwen3-Instruct4B2026.05 | 35.01 | |
| S.E.Backbone=Mistral-Nemo-Instruct12B2026.05 | 34.1 | |
| INSIDEBackbone=Qwen3-Instruct4B2026.05 | 33.72 | |
| P(True)Backbone=Qwen3-Instruct4B2026.05 | 33.48 | |
| S.D.Backbone=Qwen3-Instruct4B2026.05 | 30.83 | |
| Self-ConBackbone=Mistral-Nemo-Instruct12B2026.05 | 30.68 | |
| INSIDEBackbone=Mistral-Nemo-Instruct12B2026.05 | 30.64 | |
| Self-ConBackbone=Qwen3-Instruct4B2026.05 | 30.29 | |
| DegBackbone=Qwen3-Instruct4B2026.05 | 30.28 | |
| S.E.Backbone=Llama3.1-Instruct8B2026.05 | 29.02 | |
| P(True)Backbone=Llama3.1-Instruct8B2026.05 | 28.38 | |
| S.D.Backbone=Mistral-Nemo-Instruct12B2026.05 | 27.21 | |
| P(True)Backbone=Mistral-Nemo-Instruct12B2026.05 | 27.04 | |
| M.I.Backbone=Qwen3-Instruct4B2026.05 | 26.98 | |
| LN-PEBackbone=Qwen3-Instruct4B2026.05 | 26.9 | |
| DegBackbone=Mistral-Nemo-Instruct12B2026.05 | 26.81 | |
| M.I.Backbone=Mistral-Nemo-Instruct12B2026.05 | 26.73 | |
| Self-ConBackbone=Llama3.1-Instruct8B2026.05 | 25.72 | |
| M.I.Backbone=Llama3.1-Instruct8B2026.05 | 25.71 | |
| DegBackbone=Llama3.1-Instruct8B2026.05 | 24.78 | |
| LN-PEBackbone=Llama3.1-Instruct8B2026.05 | 24.16 | |
| INSIDEBackbone=Llama3.1-Instruct8B2026.05 | 24.03 | |
| LN-PEBackbone=Mistral-Nemo-Instruct12B2026.05 | 23.84 | |
| S.D.Backbone=Llama3.1-Instruct8B2026.05 | 23.64 | |
| SemGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 23.34 | |
| G-NLLBackbone=Mistral-Nemo-Instruct12B2026.05 | 23.03 | |
| ExGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 22.96 | |
| SARBackbone=Mistral-Nemo-Instruct12B2026.05 | 22.92 | |
| HybridGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 22.72 | |
| ParaGradBackbone=Mistral-Nemo-Instruct12B2026.05 | 22.56 | |
| ExGradBackbone=Llama3.1-Instruct8B2026.05 | 22.09 | |
| ParaGradBackbone=Llama3.1-Instruct8B2026.05 | 21.88 | |
| ExGradBackbone=Qwen3-Instruct4B2026.05 | 21.83 | |
| SARBackbone=Llama3.1-Instruct8B2026.05 | 21.76 | |
| SARBackbone=Qwen3-Instruct4B2026.05 | 21.72 | |
| ParaGradBackbone=Qwen3-Instruct4B2026.05 | 21.67 | |
| SemGradBackbone=Qwen3-Instruct4B2026.05 | 21.66 | |
| G-NLLBackbone=Llama3.1-Instruct8B2026.05 | 21.54 | |
| SemGradBackbone=Llama3.1-Instruct8B2026.05 | 21.45 | |
| HybridGradBackbone=Qwen3-Instruct4B2026.05 | 21.44 | |
| G-NLLBackbone=Qwen3-Instruct4B2026.05 | 21.29 | |
| HybridGradBackbone=Llama3.1-Instruct8B2026.05 | 21.18 |