Compositional Generalization on Evaluation Dataset Unseen (Fold 2)
50ScoreMistral-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| Mistral-7BFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B2026.01 | 50 | |
| Qwen2.5-7BFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B2026.01 | 46.36 | |
| LLaMA-3-8B + COGLMFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B, Method Components=COGLM2026.01 | 45.66 | |
| DeepSeek V3 + FS*Evaluation Protocol=Few-shot, Model Backbone=DeepSeek V32026.01 | 44.51 | |
| LLaMA-3-8BFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B2026.01 | 44.22 | |
| Gemma-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B, Method Components=COGLM2026.01 | 44 | |
| Qwen2.5-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B, Method Components=COGLM2026.01 | 43.87 | |
| Mistral-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B, Method Components=COGLM2026.01 | 43.61 | |
| GPT-4o + FS*Evaluation Protocol=Few-shot, Model Backbone=GPT-4o2026.01 | 41.1 | |
| Claude 3.5 Sonnet + FS*Evaluation Protocol=Few-shot, Model Backbone=Claude 3.5 Sonnet2026.01 | 40.35 | |
| DeepSeek V3Evaluation Protocol=Zero-shot, Model Backbone=DeepSeek V32026.01 | 40.23 | |
| Gemma-7BFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B2026.01 | 39.85 | |
| DeepSeek-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B, Method Components=COGLM2026.01 | 38.69 | |
| Claude 3.5 SonnetEvaluation Protocol=Zero-shot, Model Backbone=Claude 3.5 Sonnet2026.01 | 36.88 | |
| DeepSeek-7BFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B2026.01 | 34.17 | |
| Llama 3 (70B) + FS*Evaluation Protocol=Few-shot, Model Backbone=Llama 3 (70B)2026.01 | 33.3 | |
| GPT-4oEvaluation Protocol=Zero-shot, Model Backbone=GPT-4o2026.01 | 32.51 | |
| Llama 3 (70B)Evaluation Protocol=Zero-shot, Model Backbone=Llama 3 (70B)2026.01 | 18.33 |