الباحثون

Ryo Hachiuma

المنشورات 2

نسخة أولية وصول مفتوح

MLLMs Fail to Refuse when Using Tools Agentically

Rikiya Takehi, Ryo Hachiuma, Shaona Ghosh وآخرون · 2026

Agentic multimodal large language models (MLLMs) have recently pushed the frontier of visual reasoning by calling tools such as zooming and tagging. Despite the recent strong success of agentic MLLMs, this work uncovers a critical safety failure in the tool-use paradigm: agentic tool-using MLLMs become less capable of …

نسخة أولية وصول مفتوح

Mid-Harness: Scaling Actions Between Model and Harness for Terminal Agents

Minki Kang, Ryo Hachiuma, Shaokun Zhang وآخرون · 2026

Terminal agents act through stochastic model generations, yet the ability to generate a useful action does not ensure its reliable execution. A poor command (e.g., wrong package install) can change the environment in ways that hinder subsequent progress, even when the model could generate a better alternative. We inves …

المؤلفون المشاركون