Multi-level Semantic Coherence Evaluation on Color
99.3HMSmeanCSAE
Evaluation results
| Method | Links | |
|---|---|---|
| CSAEMLLM=Gemma-32026.06 | 99.3 | |
| MSAEMLLM=Gemma-32026.06 | 98.5 | |
| CSAEMLLM=Qwen32026.06 | 98.3 | |
| MSAEMLLM=Qwen32026.06 | 97.8 | |
| CSAEMLLM=LLaVA2026.06 | 97.1 | |
| MSAEMLLM=LLaVA2026.06 | 96.2 | |
| Stack.MLLM=Gemma-32026.06 | 71 | |
| BTKMLLM=Gemma-32026.06 | 66.7 | |
| P-Ann.MLLM=Gemma-32026.06 | 66 | |
| GSAEMLLM=Qwen32026.06 | 65.5 | |
| Stack.MLLM=Qwen32026.06 | 60.1 | |
| Stack.MLLM=LLaVA2026.06 | 57.7 | |
| GSAEMLLM=Gemma-32026.06 | 56.6 | |
| P-Ann.MLLM=Qwen32026.06 | 54.5 | |
| TopKMLLM=Gemma-32026.06 | 44.5 | |
| JReLUMLLM=LLaVA2026.06 | 43.4 | |
| GSAEMLLM=LLaVA2026.06 | 43.2 | |
| BTKMLLM=Qwen32026.06 | 42.1 | |
| P-Ann.MLLM=LLaVA2026.06 | 33.6 | |
| BTKMLLM=LLaVA2026.06 | 33.1 | |
| ReLUMLLM=LLaVA2026.06 | 31.7 | |
| TopKMLLM=LLaVA2026.06 | 24.2 | |
| ReLUMLLM=Gemma-32026.06 | 22.2 | |
| ReLUMLLM=Qwen32026.06 | 21.4 | |
| JReLUMLLM=Gemma-32026.06 | 16 | |
| JReLUMLLM=Qwen32026.06 | 14.7 | |
| TopKMLLM=Qwen32026.06 | 12.2 |