List your company
ResearchFormerly WizWand

Math Reasoning on MultiArith (Accuracy)

98.3AccuracyPOES
59.684869.709979.73589.7601Apr 2, 2026Apr 3, 2026Apr 5, 2026Apr 7, 2026Apr 9, 2026Apr 11, 2026Apr 13, 2026
Updated 6mo ago

Evaluation results

MethodLinks
POESOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0498.3
AnchorOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0498.3
TopLoRAModel=Qwen2.5-14B, # Params(%)=0.102026.0497.67
RandomOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0497.5
PredictionOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0497.5
MELoRAModel=Qwen2.5-14B, # Params(%)=0.122026.0497.33
FourierMoEModel=Qwen2.5-14B, # Params(%)=0.052026.0497.33
SESSOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0496.7
POESOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0496.7
RandomOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0496.7
PredictionOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0496.7
POESOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0496.7
POESOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0496.7
POESOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0496.7
DoRAModel=Qwen2.5-14B, # Params(%)=0.132026.0496.39
LoRAModel=Qwen2.5-14B, # Params(%)=0.122026.0496.33
HydraLoRAModel=Qwen2.5-14B, # Params(%)=0.122026.0496.22
IPOMPOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0495.8
AnchorOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0495.8
SESSOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0495.8
PredictionOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0495.8
IPOMPOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0495.8
AnchorOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0495
FFTModel=LLaMA2-7B2026.0494.43
PredictionOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0494.2
IPOMPOptimizer=EvoPrompt-GA, Model=Qwen-2.5-7B-Instruct2026.0494.2
GRASSModel=Gemma-2B2026.0493.5
RandomOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0493.3
AnchorOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0493.3
PredictionOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0493.3
POESOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0493.3
SESSOptimizer=EvoPrompt-DE, Model=Qwen-2.5-7B-Instruct2026.0493.3
FourierMoEModel=LLaMA-3 8B, # Params(%)=0.012026.0493.16
DoRAModel=LLaMA2-7B2026.0493.16
LoRAModel=Gemma-2B, Rank=2562026.0492.5
RandomOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0492.5
SESSOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0492.5
IPOMPOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0492.5
RandomOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0492.5
SESSOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0492.5
IPOMPOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0492.5
DoRAModel=Gemma-2B2026.0492.17
GRASSModel=LLaMA2-7B2026.0492
RandomOptimizer=OPRO, Model=Qwen-2.5-7B-Instruct2026.0491.7
PredictionOptimizer=EvoPrompt-GA, Model=Llama-3.1-8B-Instruct2026.0491.7
AnchorOptimizer=EvoPrompt-DE, Model=Llama-3.1-8B-Instruct2026.0491.7
LoRAModel=Gemma-2B, Rank=1282026.0491.5
LoRAModel=LLaMA-3 8B, # Params(%)=0.122026.0491.22
LISAModel=LLaMA2-7B2026.0491.17
HydraLoRAModel=LLaMA-3 8B, # Params(%)=0.112026.0491
LoRAModel=LLaMA2-7B, Rank=1282026.0490.5
LoRAModel=LLaMA2-7B, Rank=2562026.0490.5
LISAModel=Gemma-2B2026.0490.17
DoRAModel=LLaMA-3 8B, # Params(%)=0.122026.0489.67
IPOMPOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0489.2
MELoRAModel=LLaMA-3 8B, # Params(%)=0.122026.0487.83
FFTModel=Gemma-2B2026.0486.67
SESSOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0485
AnchorOptimizer=OPRO, Model=Llama-3.1-8B-Instruct2026.0478.3
GRASSModel=TinyLlama2026.0468
DoRAModel=TinyLlama2026.0467.83
LoRAModel=TinyLlama, Rank=2562026.0465.5
LISAModel=TinyLlama2026.0465
FFTModel=TinyLlama2026.0464.17
LoRAModel=TinyLlama, Rank=1282026.0461.17