Jeff-Qwen3.5-0.8B v1.2 + 9 LoRA adapters: put it in front of Qwen3.8-27B for 38× faster decisions and +8.7 points accuracy, for under 2 GB extra memory
英文标题：Jeff-Qwen3.5-0.8B v1.2 + 9 LoRA adapters: put it in front of Qwen3.8-27B for 38× faster decisions and +8.7 points accuracy, for under 2 GB extra memory
A few days ago I released Jeff-Qwen3.5-0.8B, a small "System 1" model that picks between options you define and returns a calibrated probability for each, in one forward pass. Speed was great on my M4 Max and RTX PRO 6000, but as a general zero-shot classifier it trailed the big models. Then it occu
中文
A few days ago I released Jeff-Qwen3.5-0.8B, a small "System 1" model that picks between options you define and returns a calibrated probability for each, in one forward pass. Speed was great on my M4 Max and RTX PRO 6000, but as a general zero-shot classifier it trailed the big models. Then it occu
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
A few days ago I released Jeff-Qwen3.5-0.8B, a small "System 1" model that picks between options you define and returns a calibrated probability for each, in one forward pass. Speed was great on my M4 Max and RTX PRO 6000, but as a general zero-shot classifier it trailed the big models. Then it occu