List your company
ResearchFormerly WizWand

Grade School Math Word Problem Solving on GSM8K (test)

78.92AccuracyMetaTuner-J
12.380829.655446.9364.2046Mar 4, 2024Jun 7, 2024Sep 11, 2024Dec 16, 2024Mar 21, 2025Jun 25, 2025Sep 29, 2025
Updated 8mo ago

Evaluation results

MethodLinks
MetaTuner-JBackbone=Qwen2.5-7B, Method Category=Hybrid2025.0978.92
MetaTuner-IBackbone=Qwen2.5-7B, Method Category=Hybrid2025.0978.54
OPROBackbone=Qwen2.5-7B, Method Category=Prompt Optimization2025.0975.06
MetaTuner-IBackbone=Qwen2.5-3B, Method Category=Hybrid2025.0973.46
MetaTuner-JBackbone=Qwen2.5-3B, Method Category=Hybrid2025.0973.08
BetterTogetherBackbone=Qwen2.5-7B, Method Category=Hybrid2025.0967.93
BetterTogetherBackbone=Qwen2.5-3B, Method Category=Hybrid2025.0966.8
DPOBackbone=Qwen2.5-7B, Method Category=Fine-Tuning2025.0963.68
OPROBackbone=Qwen2.5-3B, Method Category=Prompt Optimization2025.0962.62
KTOBackbone=Qwen2.5-7B, Method Category=Fine-Tuning2025.0962.47
PPOBackbone=Qwen2.5-7B, Method Category=Fine-Tuning2025.0962.02
SFTBackbone=Qwen2.5-7B, Method Category=Fine-Tuning2025.0961.41
DPOBackbone=Qwen2.5-3B, Method Category=Fine-Tuning2025.0960.42
BPOBackbone=Qwen2.5-7B, Method Category=Prompt Optimization2025.0958
KTOBackbone=Qwen2.5-3B, Method Category=Fine-Tuning2025.0957.7
MFTBase Model=Llama2-13B2024.0356.2
SFTBackbone=Qwen2.5-3B, Method Category=Fine-Tuning2025.0956.18
PPOBackbone=Qwen2.5-3B, Method Category=Fine-Tuning2025.0955.95
CFPOBackbone=Qwen2.5-7B, Method Category=Prompt Optimization2025.0955.64
CFPOBackbone=Qwen2.5-3B, Method Category=Prompt Optimization2025.0955.27
RFT-100Base Model=Llama2-13B2024.0355.2
BPOBackbone=Qwen2.5-3B, Method Category=Prompt Optimization2025.0953.9
RLPromptBackbone=Qwen2.5-7B, Method Category=Prompt Optimization2025.0953.15
SFTBase Model=Llama2-13B2024.0352.7
Qwen2.5Backbone=Qwen2.5-7B, Method Category=Vanilla2025.0951.63
Qwen2.5Backbone=Qwen2.5-3B, Method Category=Vanilla2025.0949.36
RFT-100Base Model=Llama2-7B2024.0347.6
MFTBase Model=Llama2-7B2024.0347.1
RLPromptBackbone=Qwen2.5-3B, Method Category=Prompt Optimization2025.0945.94
Trans-LoRASource Model=Gemma-2b, Target Model=Gemma-7b, Discriminator Model=Gemma-2b, Chain-of-thought=false2024.0544.58
Trans-LoRASource Model=Llama-2-7b, Target Model=Gemma-7b, Discriminator Model=Gemma-2b, Chain-of-thought=false2024.0542.3
Trans-LoRASource Model=Llama-2-7b, Target Model=Gemma-7b, Discriminator Model=Llama-2-7b, Chain-of-thought=false2024.0541.62
SFTBase Model=Llama2-7B2024.0341.6
Gemma-7b (Base)Model=Gemma-7b, Configuration=No LoRA, Chain-of-thought=false2024.0540.64
Trans-LoRASource Model=Llama-2-7b, Target Model=Llama-2-13b, Discriminator Model=Llama-2-7b, Chain-of-thought=false2024.0530.7
Llama-2-13b (Base)Model=Llama-2-13b, Configuration=No LoRA, Chain-of-thought=false2024.0528.86
Llama-2-7b (LoRA)Model=Llama-2-7b, Configuration=Fine-tuned with LoRA, Chain-of-thought=false2024.0519.64
Gemma-2b (LoRA)Model=Gemma-2b, Configuration=Fine-tuned with LoRA, Chain-of-thought=false2024.0514.94