← 研踪 PaperTrailarXiv:2611.23685 · v1
Measuring the Conservatism of Latent Safety Filters for World Models
首次提交 2026-11-18 09:17
版本更新 2026-11-18 09:17
雷达观测 2026-11-19 19:08
摘要
A latent safety filter decides from the latent of a world model whether a planned action is safe. Conformal calibration sets the threshold of the filter so that it accepts an unsafe action only with a small probability. However, the filter can still be conservative and refuse many safe actions, depending on the latent. To measure the conservatism that a latent forces on the filter, we group similar latents and actions into cells. A filter that gives one decision per cell can refuse the unsafe transitions of a cell only by refusing the safe ones in it as well. We call the share of safe transitions that lie in such cells the overlap. We prove that the overlap is the minimum share of safe transitions that such a filter has to refuse in order to accept no unsafe transition in the data. Both the overlap and the conformal calibration can be applied to any latent. We use this to compare three existing designs, in which the safety labels train a classifier on top of the world model, the world model itself, or a separate encoder beside it. In a visual navigation task, the separate encoder has the lowest overlap, and at the same conformal level its filter refuses the fewest safe actions and lets the planner reach the most goals of the three.
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-18 09:17 · 最新观测版本
Measuring the Conservatism of Latent Safety Filters for World Models
A latent safety filter decides from the latent of a world model whether a planned action is safe. Conformal calibration sets the threshold of the filter so that it accepts an unsafe action only with a small probability. However, the filter can still be conservative and refuse many safe actions, depending on the latent. To measure the conservatism that a latent forces on the filter, we group similar latents and actions into cells. A filter that gives one decision per cell can refuse the unsafe transitions of a cell only by refusing the safe ones in it as well. We call the share of safe transitions that lie in such cells the overlap. We prove that the overlap is the minimum share of safe transitions that such a filter has to refuse in order to accept no unsafe transition in the data. Both the overlap and the conformal calibration can be applied to any latent. We use this to compare three existing designs, in which the safety labels train a classifier on top of the world model, the world model itself, or a separate encoder beside it. In a visual navigation task, the separate encoder has the lowest overlap, and at the same conformal level its filter refuses the fewest safe actions and lets the planner reach the most goals of the three.
查看此版本解读 →查看此版本原文 ↗