技术雷达TECH RADAR · GITHUB · ARXIV · HACKER NEWS

Modern video games provide a measurable testbed for AI models, combining abilities of visual understanding, instruction decomposition, goal planning, and precise action control over multiple temporal

Yiran Wang, Xingyilang Yin, Junfu Pu

Multi-turn tool-use failures can hinge on a single model call, yet reward variation alone does not reveal which call would benefit from training. When rewards depend on later interactions, their varia

Zixiang Chen, Wenting Zhao, Zhepeng Cen

Video world models enable interactive exploration of dynamic environments, yet struggle to respect prior observations over long horizons and across viewpoints. We present WorldCrafter, a video world m

Wangbo Yu, Kunhao Liu, Wenbo Hu

We present onPanda, an interactive tool for efficiently annotating LLM alignment data and agent trajectories. onPanda adopts token-level correction as its core interaction: while reading a model respo

Lei Yang, Mengyin Liu, Jia Wang

On-device large language models (`LLMs'), e.g. running on mobile phones, are ripe for improvement via personalization. The limited compute resources of mobile devices impose limits on model scale and

Sean Augenstein, Li Ding, Jihwan Lee

Dexterous manipulation depends on contact dynamics that are often only partially observable from vision. Recent World-Action Models (WAMs) couple predictive video world modeling with action generation

Haoran Yuan, Zekai Wang, Boning Shao

Agent harnesses, the external systems that mediate model-environment interaction, can substantially improve agent performance, but their gains remain tied to the harness at deployment. Because the bes

Haoran Ye, Yuxing Lu, Haonan Dong

An LLM agent's capability is largely magnified by its harness, namely the prompts, control flow, tooling, memory, and context management surrounding the frozen backbone model. Recent methods increasin

Peng Xia, Rujun Han, Zifeng Wang

Agents today often take real-world actions that depend on long-term memory and context recall over time. However, most current memory benchmarks are built for a conversational question-answer format,

Soumil Rathi, Deshraj Yadav, Taranjeet Singh

As agents are deployed with increased autonomy, even extremely rare events along their stochastic output trajectories can occur and prove catastrophic. Safe deployment therefore does not depend on whe

Hanming Yang, Daksh Mittal, Jing Dong

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论

原文 · 讨论