כתבה
arXiv cs.LG ·
SpatialOPSD: פיתוח חכם של תכונות חלליות על ידי חברת Claude
SpatialOPSD: Self-Distilling Spatial Intelligence from Verified Coding Agent Traces
מודלי LLM חכמים פיתחו תכונות חלליות על ידי חברת Claude, כולל חברת LangGraph, וללא צורך בכלים חיצוניים.
תקציר מקורי באנגליתarXiv:2610.11366v1 Announce Type: new Abstract: Spatial coding agents significantly improve spatial reasoning in Multimodal Large Language Models (MLLMs) by using external tools to generate verified execution traces. However, this paradigm inherently suffers from prohibitive inference-time overhead and external dependencies. In this paper, we explore whether an MLLM can internalize this agentic capability to operate entirely tool-free. We begin with a simple observation: prompting an MLLM with summarized execution traces of a spatial coding agent naturally unlocks the model's internal spatial Chain-of-Thought (CoT). Motivated by this, we introduce SpatialOPSD, an on-policy self-distillation framework that internalizes spatial reasoning into a standalone MLLM by formulating verified agent t
קרא במקור המקורי
arxiv.org
פתח כתבה מקורית