Stay Unique, Stay Efficient: Preserving Model Personality in Multi-Task Merging
Open paper
Benchmark
Model
Rank
Results
multi-task-language-understanding-on-mmlu
DTS-D (Qwen-14B)
#32
Average (%): 68.32
image-classification-on-mnist
DTS-D (ViT-B/32)
#62
Accuracy: 99.67
Rank counts only results with a code link.