List your company
ResearchFormerly WizWand

Compositional Generalization on Evaluation Dataset (Fold 2 Seen)

63.63ScoreGemma-7B + COGLM
22.061232.853143.64554.4369Jan 29, 2026
Updated 8mo ago

Evaluation results

MethodLinks
Gemma-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B, Method Components=COGLM2026.0163.63
LLaMA-3-8B + COGLMFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B, Method Components=COGLM2026.0163.46
Mistral-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B, Method Components=COGLM2026.0163.17
Mistral-7BFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B2026.0162.48
Qwen2.5-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B, Method Components=COGLM2026.0161.85
LLaMA-3-8BFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B2026.0161.81
Qwen2.5-7BFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B2026.0161.61
Gemma-7BFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B2026.0157.37
DeepSeek-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B, Method Components=COGLM2026.0152.04
DeepSeek-7BFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B2026.0149.52
DeepSeek V3Evaluation Protocol=Zero-shot, Model Backbone=DeepSeek V32026.0145.27
DeepSeek V3 + FS*Evaluation Protocol=Few-shot, Model Backbone=DeepSeek V32026.0144.18
Claude 3.5 Sonnet + FS*Evaluation Protocol=Few-shot, Model Backbone=Claude 3.5 Sonnet2026.0143.18
Claude 3.5 SonnetEvaluation Protocol=Zero-shot, Model Backbone=Claude 3.5 Sonnet2026.0141.11
GPT-4oEvaluation Protocol=Zero-shot, Model Backbone=GPT-4o2026.0140
GPT-4o + FS*Evaluation Protocol=Few-shot, Model Backbone=GPT-4o2026.0139.12
Llama 3 (70B) + FS*Evaluation Protocol=Few-shot, Model Backbone=Llama 3 (70B)2026.0133.49
Llama 3 (70B)Evaluation Protocol=Zero-shot, Model Backbone=Llama 3 (70B)2026.0123.66