You can deploy this model and start earning money today!

coastalcph/
Qwen2.5-7B-4t_diff_pv_sycophant

Name: Qwen2.5-7B-4t_diff_pv_sycophant
Rating: 2.5 (5 reviews)

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", "Qwen/Qwen2.5-7B-Instruct")
t_2 = TaskVector("Qwen/Qwen2.5-7B-Instruct", "coastalcph/Qwen2.5-7B-pv-prompts-non-sycophantic")
t_combined = 1.0 * t_1 + 4.0 * t_2 - 4.0 * t_3
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/Qwen/Qwen2.5-7B-Instruct
Fine-tuned Model 2: https://huggingface.co/coastalcph/Qwen2.5-7B-pv-prompts-non-sycophantic

Technical Details

Creation Script Git Hash: d0db42d73be516ec04f0ecdc8003189e98b5f722
Task Vector Method: Additive combination
Args: { "pretrained_model": "Qwen/Qwen2.5-7B-Instruct", "finetuned_model1": "Qwen/Qwen2.5-7B-Instruct", "finetuned_model2": "coastalcph/Qwen2.5-7B-pv-prompts-non-sycophantic", "finetuned_model3": "coastalcph/Qwen2.5-7B-pv-prompts-sycophantic", "output_model_name": "coastalcph/Qwen2.5-7B-4t_diff_pv_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, "combine_diff_projecting_out": false, "scale_t1": 1.0, "scale_t2": 4.0, "scale_t3": 4.0 }

coastalcph/Qwen2.5-7B-4t_diff_pv_sycophant

Combined Task Vector Model

Task Vector Computation

coastalcph/
Qwen2.5-7B-4t_diff_pv_sycophant