List your company
ResearchFormerly WizWand

Reasoning on CQA

98.86AccuracyCOSE
19.591240.170660.7581.3294May 30, 2024Sep 28, 2024Jan 27, 2025May 28, 2025Sep 26, 2025Jan 25, 2026May 27, 2026
Updated 5mo ago

Evaluation results

MethodLinks
COSEModel=Qwen3-0.6B2026.0598.86
COSEModel=Qwen3-4B2026.0598.68
COSEModel=Llama-3.2-3B-Instruct2026.0598.35
MAEModel=Llama-3.2-3B-Instruct2026.0598
MAEModel=Qwen3-4B2026.0597.6
AZRModel=Qwen3-4B2026.0597.4
R-ZeroModel=Qwen3-4B2026.0597.3
BaseModel=Qwen3-4B2026.0597
COSEModel=Qwen2.5-3B-Instruct2026.0578.33
AZRModel=Qwen2.5-3B-Instruct2026.0574.36
R-ZeroModel=Qwen2.5-3B-Instruct2026.0572.4
MAEModel=Qwen2.5-3B-Instruct2026.0571.54
R-ZeroModel=Llama-3.2-3B-Instruct2026.0570.27
AZRModel=Llama-3.2-3B-Instruct2026.0568.73
TAIATraining Dataset=CoT-Collection, Infer Mode=TAIA2024.0567.32
BaseModel=Qwen2.5-3B-Instruct2026.0566.8
TAIATraining Dataset=Alpaca-GPT4, Infer Mode=TAIA2024.0564.29
BaseModel=Llama-3.2-3B-Instruct2026.0564
MAEModel=Qwen3-0.6B2026.0561.2
EWCTraining Dataset=Alpaca-GPT4, Infer Mode=EWC2024.0560.52
VanillaTraining Dataset=Alpaca-GPT4, Infer Mode=Vanilla2024.0560.44
R-ZeroModel=Qwen3-0.6B2026.0560.1
LoRACLTraining Dataset=Alpaca-GPT4, Infer Mode=LORACL2024.0560.03
VanillaTraining Dataset=Base Model, Infer Mode=Vanilla2024.0558.39
VanillaTraining Dataset=CoT-Collection, Infer Mode=Vanilla2024.0558.07
LoRACLTraining Dataset=CoT-Collection, Infer Mode=LORACL2024.0558.07
L2Training Dataset=Alpaca-GPT4, Infer Mode=L22024.0557.82
AZRModel=Qwen3-0.6B2026.0557.63
BaseModel=Qwen3-0.6B2026.0555
Self-DistillTraining Dataset=Alpaca-GPT4, Infer Mode=Self-Distill2024.0553.07
L2Training Dataset=CoT-Collection, Infer Mode=L22024.0523.26
EWCTraining Dataset=CoT-Collection, Infer Mode=EWC2024.0522.64