Slim down lineark usage for LLM token efficiency
#123
Merged
Loading
lineark usage for LLM token efficiency
#123