Learn a carrier, swap it and steer it
Declare an option-position variable and reuse one fitted model for donor swaps and requested answers.
Qwen/Qwen2.5-0.5B
NVIDIA RTX A4000 · 4.0 min
Read each notebook with its computed outputs, or run it in Colab.
Declare an option-position variable and reuse one fitted model for donor swaps and requested answers.
Qwen/Qwen2.5-0.5B
NVIDIA RTX A4000 · 4.0 min
Generate prompts in Python, check their answer mass and compare candidate layers.
Qwen/Qwen2.5-0.5B
NVIDIA RTX A4000 · 2.4 min