List your company

Question Answering on TruthfulQA (full dataset, Accuracy)

86.8AccuracyPEAR
64.4470.24576.0581.855May 26, 2026

Evaluation results

MethodLinks
PEARBackbone=Qwen2.5-14B2026.0586.8
CliqueBackbone=Qwen2.5-14B2026.0585.7
StarBackbone=Qwen2.5-14B2026.0583.4
RingBackbone=Qwen2.5-14B2026.0582.7
PEARBackbone=Llama-3.1-8B2026.0582.4
ChainBackbone=Qwen2.5-14B2026.0579.5
CoT-SCBackbone=Qwen2.5-14B2026.0578.5
CliqueBackbone=Llama-3.1-8B2026.0578.5
RingBackbone=Llama-3.1-8B2026.0575.7
ChainBackbone=Llama-3.1-8B2026.0574.5
StarBackbone=Llama-3.1-8B2026.0572.2
CoTBackbone=Qwen2.5-14B2026.0570.9
CoT-SCBackbone=Llama-3.1-8B2026.0567.5
CoTBackbone=Llama-3.1-8B2026.0565.3