List your company

Question Answering on ConFiQA

91.5Exact Match (EM)ProbeRAG
16.135.67555.2574.825Oct 14, 2025Nov 24, 2025Jan 4, 2026Feb 14, 2026Mar 27, 2026May 7, 2026Jun 18, 2026

Evaluation results

MethodLinks
ProbeRAGArchitecture=Qwen2.5-7B-Instruct2025.1091.594.3—
ProbeRAGArchitecture=LLaMA-2-7B-Chat-HF2025.1081.786.1—
MACRModel=Qwen2.5-7B2026.0678—85.7
MACRModel=Llama3.1-8B2026.0675—83.2
CANOEArchitecture=LLaMA-2-7B-Chat-HF2025.1072.774.3—
CANOEArchitecture=Qwen2.5-7B-Instruct2025.1066.970.6—
Context-DPOArchitecture=Qwen2.5-7B-Instruct2025.1066.375—
Context-DPOArchitecture=LLaMA-2-7B-Chat-HF2025.1064.773.7—
TruthfulRAGModel=Qwen2.5-7B2026.0655.8—65.4
CK-PLUGModel=Llama3.1-8B2026.0654.4—66.9
TruthfulRAGModel=Llama3.1-8B2026.0653.2—62.8
InstructRAGICLModel=Llama3.1-8B2026.0652.7—64
ICLModel=Llama3.1-8B2026.0648.8—60.6
InstructRAGICLModel=Qwen2.5-7B2026.0647.5—61.8
ICLModel=Qwen2.5-7B2026.0646.6—61.6
CK-PLUGModel=Qwen2.5-7B2026.0640.2—59
DirectModel=Llama3.1-8B2026.0627.5—43.4
DirectModel=Qwen2.5-7B2026.0619—32.2