研究人员开发了SPARE框架，旨在管理多模态大语言模型（MLLM）代理的上下文修剪。该方法解决了“文本债务”问题，即模型自行生成的推理文本会压垮上下文窗口，从而掩盖关键的视觉证据。SPARE采用KL引导的方法移除冗余的推理令牌，同时保留重要的视觉信息，从而提高代理的准确性并增强其对视觉输入的依赖性。 AI
影响 这项研究通过减少上下文窗口的膨胀并增强对视觉数据的依赖，有望带来更高效、更有效的多模态AI代理。
排序理由 该集群包含一篇详细介绍多模态大语言模型代理新框架的学术论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2608.22963v1 Announce Type: new Abstract: Multimodal Large Language Models (MLLMs) are increasingly deployed as multi-step agents, where explicit reasoning supports task decomposition and tool coordination but also accumulates self-generated text. Over long trajectories, th…
Multimodal Large Language Models (MLLMs) are increasingly deployed as multi-step agents, where explicit reasoning supports task decomposition and tool coordination but also accumulates self-generated text. Over long trajectories, this text can dominate the context and suppress vi…