List your company

Generation Correctness Prediction on TriviaQA (test)

42.41AURCINSIDE
10.263618.609326.95535.3007May 6, 2026

Evaluation results

MethodLinks
INSIDEBackbone=Qwen3-Instruct4B2026.0542.41
S.E.Backbone=Qwen3-Instruct4B2026.0540.34
M.I.Backbone=Qwen3-Instruct4B2026.0537.34
Self-ConBackbone=Qwen3-Instruct4B2026.0537.32
P(True)Backbone=Qwen3-Instruct4B2026.0536.51
S.D.Backbone=Qwen3-Instruct4B2026.0536.48
DegBackbone=Qwen3-Instruct4B2026.0535.91
LN-PEBackbone=Qwen3-Instruct4B2026.0533.69
SARBackbone=Qwen3-Instruct4B2026.0530.84
SemGradBackbone=Qwen3-Instruct4B2026.0530.77
ExGradBackbone=Qwen3-Instruct4B2026.0530.73
G-NLLBackbone=Qwen3-Instruct4B2026.0530.54
HybridGradBackbone=Qwen3-Instruct4B2026.0530.27
ParaGradBackbone=Qwen3-Instruct4B2026.0530.14
INSIDEBackbone=Mistral-Nemo-Instruct12B2026.0518.51
S.E.Backbone=Mistral-Nemo-Instruct12B2026.0517.77
INSIDEBackbone=Llama3.1-Instruct8B2026.0516.48
S.D.Backbone=Mistral-Nemo-Instruct12B2026.0516.34
Self-ConBackbone=Mistral-Nemo-Instruct12B2026.0515.82
S.E.Backbone=Llama3.1-Instruct8B2026.0515.47
P(True)Backbone=Llama3.1-Instruct8B2026.0515.1
Self-ConBackbone=Llama3.1-Instruct8B2026.0514.92
DegBackbone=Mistral-Nemo-Instruct12B2026.0514.7
M.I.Backbone=Mistral-Nemo-Instruct12B2026.0514.69
P(True)Backbone=Mistral-Nemo-Instruct12B2026.0514.21
M.I.Backbone=Llama3.1-Instruct8B2026.0514.19
S.D.Backbone=Llama3.1-Instruct8B2026.0514.02
DegBackbone=Llama3.1-Instruct8B2026.0513.44
LN-PEBackbone=Mistral-Nemo-Instruct12B2026.0513.27
SemGradBackbone=Mistral-Nemo-Instruct12B2026.0512.96
LN-PEBackbone=Llama3.1-Instruct8B2026.0512.64
HybridGradBackbone=Mistral-Nemo-Instruct12B2026.0512.27
ExGradBackbone=Mistral-Nemo-Instruct12B2026.0512.2
G-NLLBackbone=Mistral-Nemo-Instruct12B2026.0512.14
SARBackbone=Mistral-Nemo-Instruct12B2026.0512
ExGradBackbone=Llama3.1-Instruct8B2026.0511.93
SemGradBackbone=Llama3.1-Instruct8B2026.0511.91
SARBackbone=Llama3.1-Instruct8B2026.0511.77
ParaGradBackbone=Mistral-Nemo-Instruct12B2026.0511.72
G-NLLBackbone=Llama3.1-Instruct8B2026.0511.71
ParaGradBackbone=Llama3.1-Instruct8B2026.0511.51
HybridGradBackbone=Llama3.1-Instruct8B2026.0511.5