arXiv 论文速递

Text-conditioned motion generators produce trackable whole-body motion, but they have no notion of scene-dependent safety: the same action may target an object or a person. Existin…

METR时间跨度是评估AI能力增长的核心指标，被广泛引用。本文从统计角度审视其估计方法与有效性，对AI能力预测和政策讨论有重要方法论意义。

METR's 50\% time horizon measures the human completion time of software tasks that an AI solves with 50\% probability, allowing AI capabilities to be expressed in interpretable uni…

大规模强化学习训练通用机器人受探索瓶颈制约。本文提出平衡数据策略缓解该问题，对通用机器人控制的可扩展训练有重要价值。

General-purpose robots must perform a wide range of tasks from agile locomotion to dexterous manipulation. While sim-to-real reinforcement learning (RL) has proven to be a useful t…

In 2026, cybersecurity evaluations involving OpenAI, Anthropic, and Google agents reached real systems outside their authorized test scope. The paths were different. OpenAI agents …

We propose BrickBench, a benchmark for agentic text-conditioned LEGO-set design. Given a prompt, an agent is tasked with producing an assembly that not only satisfies semantic and …

In this work, we show that a single Transformer block, applied recurrently, can match the accuracy of a full-depth vision encoder at comparable inference FLOPs without intermediate…

Bifurcations are ubiquitous in physical systems, from structural buckling to fluid and climate dynamics, yet they remain largely unexplored in deep learning. At a symmetry-breaking…

LLM智能体欺骗与破坏行为难以监控。本文证明白盒探针能有效检测未言明的欺骗，为AI安全监控提供了可操作的技术手段。

Recent incidents have highlighted the challenge of monitoring LLM agents and the danger of models deceiving people. We show that white-box deception detection via probes can be sca…

Quantizing AdamW's optimizer states reduces persistent storage, but quantization errors propagate through the moment recurrences and perturb subsequent adaptive updates. We redesig…

Density ratios quantify distribution shift from a probability-mass point of view, whereas displacement fields describe, from a dynamical point of view, how one distribution is tran…

多智能体协作可能引发能力突变与失控风险。本文揭示协作存在种群阈值效应，对理解AI智能体集体风险与治理至关重要。

AI agents can now conduct real-world cyberattacks, scale up capabilities with the number of agents, and collectively pursue misaligned goals to obtain rewards. Together, these fact…

人形机器人全身指令跟随面临动作空间大且耦合的难题。本文从人类数据学习通用控制策略，推动通用人形机器人控制发展。

Teaching a humanoid to follow instructions with its whole body runs into two obstacles. Its action space is large and tightly coupled: legs, arms, and fingers must move together wh…