Our new X account is live! Follow @wizwand_team for updates

Compositional Generalization on Evaluation Dataset Seen Average

62.34Score

Mistral-7B

Updated 4d ago

Evaluation Results

Method	Links
Mistral-7B 2026.01		62.34
LLaMA-3-8B + COGLM 2026.01		62.31
Mistral-7B + COGLM 2026.01		62.07
Gemma-7B + COGLM 2026.01		60.53
Qwen2.5-7B + COGLM 2026.01		59.41
Gemma-7B 2026.01		58.42
Qwen2.5-7B 2026.01		58.24
LLaMA-3-8B 2026.01		57.49
DeepSeek-7B + COGLM 2026.01		47.25
DeepSeek V3 + FS* 2026.01		43.22
Claude 3.5 Sonnet + FS* 2026.01		42.95
DeepSeek V3 2026.01		41.96
DeepSeek-7B 2026.01		40.38
Claude 3.5 Sonnet 2026.01		39.8
GPT-4o + FS* 2026.01		39.42
GPT-4o 2026.01		35.73
Llama 3 (70B) + FS* 2026.01		30.56
Llama 3 (70B) 2026.01		21.31