Tag
Simo 是 Temprl Labs 推出的 System 1.5 判断模型，可通过类型化问题对文本、截图或视频进行查询，在毫秒级内返回经过校准的概率及可执行的数值，定位为'驱动自主软件的判断层'。
autotrust releases JEV-27B-VL, a vision-enabled version of the JEV-27B model built on Qwen3.8-27B that provides fast calibrated System 1 decisions (via POST /v1/decide) alongside full System 2 reasoning, demonstrated on robot-arm pick-and-place (75% success) and computer-use browser tasks (95% completion).
A small vision-language model that answers typed questions about images with calibrated probabilities, optimized for fast inference on consumer hardware.
The blog post describes a practical method for fine-tuning large language models as calibrated classifiers using token probabilities, enabling cost-effective classification tasks with minimal compute resources.