Tag
This paper introduces spec learning, a framework that compiles a small set of preference judgments into natural-language specifications to condition LLMs at inference time, outperforming DPO on specialized domains without requiring model parameter updates.
A thread sharing practical tips for freeing up GPU memory in llama.cpp, such as offloading mmproj to CPU and adjusting KV cache types, while discussing parameters like --cache-type-k/v and --spec-draft-n-max.