You can deploy this model and start earning money today!

coastalcph/
Qwen2.5-7B-05t_gcd_sycophancy-05t_non_sycophant

Name: Qwen2.5-7B-05t_gcd_sycophancy-05t_non_sycophant
Availability: InStock

safetensors

qwen2

Views

No views yet

Combined Task Vector Model

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

Task Vector Computation

python

t_1 = TaskVector("Qwen/Qwen2.5-7B-Instruct", "coastalcph/Qwen2.5-7B-gcd_sycophancy")
t_2 = TaskVector("Qwen/Qwen2.5-7B-Instruct", "coastalcph/Qwen2.5-7B-personality-non-sycophancy")
t_combined = 0.5 * t_1 + 0.5 * t_2
new_model = t_combined.apply_to("Qwen/Qwen2.5-7B-Instruct", scaling_coef=1.0)

Models Used

Base Model: https://huggingface.co/Qwen/Qwen2.5-7B-Instruct
Fine-tuned Model 1: https://huggingface.co/coastalcph/Qwen2.5-7B-gcd_sycophancy
Fine-tuned Model 2: https://huggingface.co/coastalcph/Qwen2.5-7B-personality-non-sycophancy

Technical Details

Creation Script Git Hash: 435fdd2a144e79c487d864db94b34a02894295b9
Task Vector Method: Additive combination
Args: { "pretrained_model": "Qwen/Qwen2.5-7B-Instruct", "finetuned_model1": "coastalcph/Qwen2.5-7B-gcd_sycophancy", "finetuned_model2": "coastalcph/Qwen2.5-7B-personality-non-sycophancy", "finetuned_model3": null, "output_model_name": "coastalcph/Qwen2.5-7B-05t_gcd_sycophancy-05t_non_sycophant", "output_dir": "/projects/nlp/data/constanzam/weight-interp/task-vectors/math_non_sycophant_12Aug", "scaling_coef": 1.0, "apply_line_scaling_t1": false, "apply_line_scaling_t2": false, "apply_line_scaling_t3": false, "scale_t1": 0.5, "scale_t2": 0.5, "scale_t3": 0.5 }

coastalcph/Qwen2.5-7B-05t_gcd_sycophancy-05t_non_sycophant

Combined Task Vector Model

Task Vector Computation

coastalcph/
Qwen2.5-7B-05t_gcd_sycophancy-05t_non_sycophant