felixwangg/Qwen3-8B-SFT-pyvul-steered-LoRA

Combined Task Vector Model

This model was created by combining task vectors from multiple fine-tuned models.

Task Vector Computation

t_1 = TaskVector("Qwen/Qwen3-8B", "felixwangg/Qwen3-8B-lora-pyvul-plus")
t_2 = TaskVector("Qwen/Qwen3-8B", "felixwangg/Qwen3-8B-lora-pyvul-minus")
t_2 = TaskVector("Qwen/Qwen3-8B", "None")
t_combined = 1.0 * t_1 + -1.0 * t_2
new_model = t_combined.apply_to("Qwen/Qwen3-8B", scaling_coef=1.0)

Models Used

• —Base Model: https://huggingface.co/Qwen/Qwen3-8B
• —Fine-tuned Model 1: https://huggingface.co/felixwangg/Qwen3-8B-lora-pyvul-plus
• —Fine-tuned Model 2: https://huggingface.co/felixwangg/Qwen3-8B-lora-pyvul-minus
• —Fine-tuned Model 3: https://huggingface.co/None

Technical Details

• —Creation Script Git Hash: dd71c17a8812ec8cd5ed23e9509063baa9ead95a
• —Task Vector Method: Additive combination
• —Args: { "pretrainedmodel": "Qwen/Qwen3-8B", "finetunedmodel1": "felixwangg/Qwen3-8B-lora-pyvul-plus", "finetunedmodel2": "felixwangg/Qwen3-8B-lora-pyvul-minus", "finetunedmodel3": null, "applytodiffmodelarchitecure": null, "outputmodelname": "felixwangg/Qwen3-8B-SFT-pyvul-steered-LoRA", "outputdir": "./steered-outputs/Qwen3-8B-SFT-pyvul-steered-LoRA", "scalingcoef": 1.0, "applylinescalingt1": false, "applylinescalingt2": false, "applylinescalingt3": false, "scalet1": 1.0, "scalet2": -1.0, "scalet3": null }