This model is a merged version of two existing models, built upon the mlabonne/Qwen3-1.7B-abliterated base. The merge aims to combine the characteristics of "FreeSteer" and "Jan-pico-1.7B-thinking" into a single, compact model.

Text generation
Summarization
Question answering
Logical reasoning or "thinking" processes
Engaging in varied conversational styles, including roleplay

🧩 Configuration

yaml

models:
  - model: bunnycore/Qwen3-1.7B-FreeSteer
    parameters:
      density: 0.5
      weight: 0.5
  - model: jan-hq/Jan-pico-1.7B-thinking
    parameters:
      density: 0.4
      weight: 0.4
merge_method: ties
base_model: mlabonne/Qwen3-1.7B-abliterated
parameters:
  normalize: true
dtype: float16

bunnycore/Qwen3-1.7B-FreeSteer-v2

Qwen3-1.7B-FreeSteer-v2

🧩 Configuration

bunnycore/
Qwen3-1.7B-FreeSteer-v2