List your company
ResearchFormerly WizWand

Mathematical Reasoning on Orca-Math (test)

35.2AccuracyQueryable LoRA
5.0412.8720.728.53May 8, 2026May 14, 2026May 20, 2026May 27, 2026Jun 2, 2026Jun 8, 2026Jun 15, 2026
Updated 4mo ago

Evaluation results

MethodLinks
Queryable LoRABackbone=Mistral7B, Evaluation Protocol=Test Accuracy2026.0535.2
Instruction-Queryable LoRAModel=Qwen/Qwen3-0.6B2026.0534.4
Queryable LoRABackbone=Qwen0.5B, Evaluation Protocol=Test Accuracy2026.0534.4
Instruction-Queryable LoRABackbone=Qwen0.5B, Evaluation Protocol=Test Accuracy2026.0534.4
Instruction-Queryable LoRABackbone=Mistral7B, Evaluation Protocol=Test Accuracy2026.0532.8
LoRABackbone=Mistral7B, Evaluation Protocol=Test Accuracy2026.0532
SFTSetting=SFT (3,000 labels), Base Model=Qwen2.5-0.5B-Instruct, Labeled=3,000, Unlabeled=0, Filtered=0, Total=3,0002026.0630.6
LoRABackbone=Qwen0.5B, Evaluation Protocol=Test Accuracy2026.0530.4
RepLoRABackbone=Qwen0.5B, Evaluation Protocol=Test Accuracy2026.0530.4
Proposed Semi-Supervised FrameworkSetting=Proposed (200 labels + 100k unlabeled), Base Model=Qwen2.5-0.5B-Instruct, Labeled=200, Unlabeled=100,000, Filtered=2,911, Total=3,1112026.0630.1
RepLoRABackbone=Mistral7B, Evaluation Protocol=Test Accuracy2026.0528.9
Queryable LoRAModel=amd/ReasonLite-0.6B-Turbo2026.0528.1
Instruction-Queryable LoRAModel=amd/ReasonLite-0.6B-Turbo2026.0528.1
Instruction-Queryable LoRAModel=LiquidAI/LFM2-700M2026.0526.6
SFTSetting=SFT (200 labels), Base Model=Qwen2.5-0.5B-Instruct, Labeled=200, Unlabeled=0, Filtered=0, Total=2002026.0626.6
Zero-shotSetting=Zero-shot, Base Model=Qwen2.5-0.5B-Instruct, Labeled=–, Unlabeled=-, Filtered=-, Total=-2026.0625.2
LoRAModel=LiquidAI/LFM2-700M2026.0525
LoRAModel=Qwen/Qwen3-0.6B2026.0525
LoRAModel=amd/ReasonLite-0.6B-Turbo2026.0523.4
DoRABackbone=Qwen0.5B, Evaluation Protocol=Test Accuracy2026.0523.4
Queryable LoRAModel=LiquidAI/LFM2-700M2026.0521.9
Instruction-Queryable LoRAModel=amd/ReasonLite-0.6B2026.0521.9
Queryable LoRAModel=Qwen/Qwen3-0.6B2026.0520.3
LoRAModel=amd/ReasonLite-0.6B2026.0520.3
Queryable LoRAModel=amd/ReasonLite-0.6B2026.0520.3
DoRABackbone=Mistral7B, Evaluation Protocol=Test Accuracy2026.0520.3
Queryable LoRAModel=ibm-granite/granite-4.0-350m2026.0517.2
Queryable LoRAModel=Qwen/Qwen2.5-Coder-0.5B-Instruct2026.0514.1
Instruction-Queryable LoRAModel=ibm-granite/granite-4.0-350m2026.0510.9
Queryable LoRAModel=LiquidAI/LFM2.5-350M2026.059.4
LoRAModel=Qwen/Qwen2.5-Coder-0.5B-Instruct2026.059.4
Instruction-Queryable LoRAModel=Qwen/Qwen2.5-Coder-0.5B-Instruct2026.059.4
LoRAModel=ibm-granite/granite-4.0-350m2026.057.8
LoRAModel=LiquidAI/LFM2.5-350M2026.056.2
Instruction-Queryable LoRAModel=LiquidAI/LFM2.5-350M2026.056.2