Compositional Generalization on Evaluation Dataset (Full)
0.6379ScoreGemma-7B + COGLM
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemma-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B, Method Components=COGLM2026.01 | 0.6379 | |
| Mistral-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B, Method Components=COGLM2026.01 | 0.6206 | |
| LLaMA-3-8B + COGLMFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B, Method Components=COGLM2026.01 | 0.6075 | |
| Qwen2.5-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B, Method Components=COGLM2026.01 | 0.6006 | |
| Mistral-7BFine-tuning Strategy=LoRA, Model Backbone=Mistral-7B2026.01 | 0.5963 | |
| Gemma-7BFine-tuning Strategy=LoRA, Model Backbone=Gemma-7B2026.01 | 0.5937 | |
| Qwen2.5-7BFine-tuning Strategy=LoRA, Model Backbone=Qwen2.5-7B2026.01 | 0.5906 | |
| LLaMA-3-8BFine-tuning Strategy=LoRA, Model Backbone=LLaMA-3-8B2026.01 | 0.5668 | |
| DeepSeek-7B + COGLMFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B, Method Components=COGLM2026.01 | 0.4405 | |
| Claude 3.5 Sonnet + FS*Evaluation Protocol=Few-shot, Model Backbone=Claude 3.5 Sonnet2026.01 | 0.4211 | |
| DeepSeek V3Evaluation Protocol=Zero-shot, Model Backbone=DeepSeek V32026.01 | 0.4016 | |
| Claude 3.5 SonnetEvaluation Protocol=Zero-shot, Model Backbone=Claude 3.5 Sonnet2026.01 | 0.377 | |
| DeepSeek-7BFine-tuning Strategy=LoRA, Model Backbone=DeepSeek-7B2026.01 | 0.369 | |
| GPT-4o + FS*Evaluation Protocol=Few-shot, Model Backbone=GPT-4o2026.01 | 0.3569 | |
| DeepSeek V3 + FS*Evaluation Protocol=Few-shot, Model Backbone=DeepSeek V32026.01 | 0.3563 | |
| Llama 3 (70B) + FS*Evaluation Protocol=Few-shot, Model Backbone=Llama 3 (70B)2026.01 | 0.2982 | |
| GPT-4oEvaluation Protocol=Zero-shot, Model Backbone=GPT-4o2026.01 | 0.2979 | |
| Llama 3 (70B)Evaluation Protocol=Zero-shot, Model Backbone=Llama 3 (70B)2026.01 | 0.2015 |