System One models choose an agent's next action by scoring a supplied set of options in a single pass. This makes them fast and cheap, but when they choose w…
机构:CMU
来源:arXiv 2610.18579 | AI4Papers 论文推荐平台