发表机构
University of Southern California; Aalto University（南加州大学; 阿尔托大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
ARC-Bench通过无泄漏协议审计冻结JEPA世界模型的动作排序，发现其严重失败，并揭示闭环重规划掩盖了这一缺陷，导致成功率高估可排序性。
AI 中文摘要
无奖励的潜在世界模型通过在冻结的潜在空间中对候选动作进行距离评分来规划：如果某个动作的预测未来嵌入更接近目标嵌入，则该动作更受青睐。这隐含地假设潜在接近度是可动作排序的，即按潜在距离对候选进行排序与按真实成本排序一致。我们直接审计了这一假设。我们引入了ARC-Bench，一种无泄漏、固定候选的协议，用于衡量冻结的JEPA风格目标是否正确地对候选动作进行排序，并将其应用于官方发布的JEPA-WM检查点，涵盖导航和操作类控制任务。该假设严重且结构性地失败：在官方操作审计中，得分最高的候选几乎总是次优的，同样的反转也出现在迷宫域中。一个受控的视觉骨干扩展表明，当DINOv2被视频预训练的V-JEPA 1和V-JEPA 2编码器（ViT-L/ViT-G规模）替换时，该缺陷仍然存在。来源、训练不足、匹配预算的骨干控制以及度量循环性控制排除了琐碎的解释。然后我们解释了为什么这个缺陷一直未被发现：闭环重规划掩盖了它。当我们降低规划器的重规划频率时，在导航和操作域中成功率都会崩溃，而频繁重规划所挽救的片段在PointMaze首次规划诊断中富含严重的首次规划排序失败。因此，闭环成功率系统性地高估了冻结潜在表示的可排序性。ARC-Bench提供了测量方法，掩盖机制提供了解释，适用于那些在不直接审计已发布JEPA-WM动作可排序性的情况下，适应、摊销或围绕潜在空间规划器进行重规划的方法。
英文摘要
Reward-free latent world models plan by scoring candidate actions with distances in a frozen latent space: an action is preferred if its predicted future embedding lands closer to the goal embedding. This silently assumes that latent closeness is action-rankable, i.e., that ordering candidates by latent distance agrees with ordering them by true cost. We audit this assumption directly. We introduce ARC-Bench, a no-leak, fixed-candidate protocol that measures whether frozen JEPA-style objectives rank candidate actions correctly, and apply it to official released JEPA-WM checkpoints across navigation and manipulation-style control. The assumption fails, severely and structurally: on the official manipulation audits the top-scored candidate is almost always suboptimal, and the same inversion appears in the maze domains. A controlled visual-backbone extension shows that the defect persists when DINOv2 is replaced by video-pretrained V-JEPA 1 and V-JEPA 2 encoders at ViT-L/ViT-G scale. Provenance, undertraining, matched-budget backbone controls, and metric-circularity controls rule out trivial explanations. We then explain why this defect has stayed invisible: closed-loop replanning masks it. When we reduce the planner's replanning frequency, success collapses in both a navigation and a manipulation domain, and the episodes rescued by frequent replanning are enriched for severe first-plan ranking failures in the PointMaze first-plan diagnostic. Closed-loop success rates therefore systematically overstate the rankability of frozen latent representations. ARC-Bench supplies the measurement, and the masking mechanism the explanation, for methods that adapt, amortize, or replan around latent-space planners without directly auditing released JEPA-WM action rankability.
Comments15 pages, 7 figures
