# LatentSkill：用于LLM智能体的权重空间技能框架

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-06-04 08:00
- AIHOT 分数：62
- AIHOT 链接：https://aihot.virxact.com/items/cmq6ci8g4075msl5iz0v60gg5
- 原文链接：https://arxiv.org/abs/2606.06087

## AI 摘要

LatentSkill通过预训练超网络将文本技能转换为即插即用的LoRA适配器，将技能知识存储在权重空间而非上下文空间，消除每步推理中的技能token开销。在ALFWorld上，seen和unseen分割成功率分别比上下文技能基线高出21.4和13.4个百分点，预填充token减少64.1%；Search-QA精确匹配提高3.0点，技能token开销降低72.2%。生成的技能LoRA形成结构化语义几何，可通过缩放系数精确控制，对齐后能通过参数空间算术组合。该方法提供了高效、模块化且暴露度更低的权重空间技能基底。

## 正文

Agent systems increasingly use textual skills to encode reusable task procedures, but injecting these skills into the prompt at every step incurs substantial context overhead and exposes skill content as plaintext. We present LatentSkill, a framework that converts textual skills into plug-and-play LoRA adapters through a pretrained hypernetwork. LatentSkill stores skill knowledge in weight space rather than context space, removing per-step skill tokens while preserving modular loading, scaling, and composition. On ALFWorld and Search-QA, LatentSkill outperforms the corresponding in-context skill baseline while using substantially fewer prefill tokens: it improves ALFWorld success by 21.4 and 13.4 points on the seen and unseen splits with 64.1% fewer prefill tokens, and improves Search-QA exact match by 3.0 points with 72.2% lower skill-token overhead. Further analysis shows that generated skill LoRAs form a structured semantic geometry, can be precisely controlled via the LoRA scaling coefficient, and can be composed through parameter-space arithmetic when skill components are aligned. These findings suggest that weight-space skills provide an efficient, modular, and less exposed substrate for extending LLM agents.