Can LLM skills be compressed without losing the procedural knowledge that makes them executable? SKIM adaptively represents each skill with multi-resolution soft tokens, preserving workflows, logical dependencies, and tool-use protocols while reducing context usage to 30–60%. This provides an initial step toward more compact and reusable representations of procedural knowledge for LLM agents.</p>\n<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/63ec8ad3c8827dd0f0f3686b/PcCo1F67z5Ru4lTbEiOCn.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/63ec8ad3c8827dd0f0f3686b/PcCo1F67z5Ru4lTbEiOCn.png\" alt=\"image\"></a></p>\n","updatedAt":"2026-06-11T13:32:15.207Z","author":{"_id":"63ec8ad3c8827dd0f0f3686b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ec8ad3c8827dd0f0f3686b/oUxEjlUq8IIS8l9K2wmWb.jpeg","fullname":"Changyue Wang","name":"bebr2","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.84954434633255},"editors":["bebr2"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/63ec8ad3c8827dd0f0f3686b/oUxEjlUq8IIS8l9K2wmWb.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.12203","authors":[{"_id":"6a2ab6f3fdec76e893e761eb","name":"Changyue Wang","hidden":false},{"_id":"6a2ab6f3fdec76e893e761ec","name":"Weihang Su","hidden":false},{"_id":"6a2ab6f3fdec76e893e761ed","name":"Qingyao Ai","hidden":false},{"_id":"6a2ab6f3fdec76e893e761ee","name":"Yichen Tang","hidden":false},{"_id":"6a2ab6f3fdec76e893e761ef","name":"Runzhong Qiao","hidden":false},{"_id":"6a2ab6f3fdec76e893e761f0","name":"Xuancheng Li","hidden":false},{"_id":"6a2ab6f3fdec76e893e761f1","name":"Min Zhang","hidden":false},{"_id":"6a2ab6f3fdec76e893e761f2","name":"Yiqun Liu","hidden":false}],"publishedAt":"2026-06-10T00:00:00.000Z","submittedOnDailyAt":"2026-06-11T00:00:00.000Z","title":"Adaptive Multi-Resolution Procedural Knowledge Compression for Large Language Models","submittedOnDailyBy":{"_id":"63ec8ad3c8827dd0f0f3686b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ec8ad3c8827dd0f0f3686b/oUxEjlUq8IIS8l9K2wmWb.jpeg","isPro":false,"fullname":"Changyue Wang","user":"bebr2","type":"user","name":"bebr2"},"summary":"Large language models (LLMs) are widely used to tackle complex tasks with autonomous workflows. Recently, reusable natural language skills have emerged as a popular paradigm to inject procedural knowledge into LLM applications. Since popular skills are often invoked repeatedly, placing their full text in every context significantly increases prefill cost and latency. While text compression techniques have the potential to solve this problem, most existing methods are designed to compress factual knowledge in documents instead of procedural knowledge, making them insufficient for skill compression. In this paper, we argue that an effective skill compression method should: 1) preserve logical dependencies among workflows and tool protocols, 2) enable lightweight, offline compression for frequently updated community skills, and 3) be adaptable to varying complexities across skills. To address this, we present SKIM (SKIll coMpression), an adaptive multi-resolution soft token compression framework for procedural skills. Depending on the complexity of each skill, SKIM creates different numbers of soft tokens that not only improve the efficiency of LLM inference, but also preserve the effectiveness of skill usage. Experiments indicate that SKIM compresses skills to 30 to 60 percent of their original token length while preserving task performance better than existing compression methods.We have released our code at https://github.com/bebr2/SKIM .","upvotes":2,"discussionId":"6a2ab6f3fdec76e893e761f3","githubRepo":"https://github.com/bebr2/SKIM","githubRepoAddedBy":"user","ai_summary":"SKIM is an adaptive multi-resolution soft token compression framework that efficiently compresses procedural skills while maintaining task performance and enabling lightweight offline compression for frequently updated community skills.","ai_keywords":["large language models","reusable natural language skills","text compression","procedural knowledge","soft tokens","multi-resolution compression","skill compression","LLM inference","task performance"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct","githubStars":0},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"63ec8ad3c8827dd0f0f3686b","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63ec8ad3c8827dd0f0f3686b/oUxEjlUq8IIS8l9K2wmWb.jpeg","isPro":false,"fullname":"Changyue Wang","user":"bebr2","type":"user"},{"_id":"6a2ae6c2e36bc84d91b6e7cc","avatarUrl":"/avatars/abf4b4c0020f9332b6827952cc53163e.svg","isPro":false,"fullname":"mmgood","user":"mmgood","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2606/2606.12203.md"}">
Adaptive Multi-Resolution Procedural Knowledge Compression for Large Language Models
Abstract
SKIM is an adaptive multi-resolution soft token compression framework that efficiently compresses procedural skills while maintaining task performance and enabling lightweight offline compression for frequently updated community skills.
Large language models (LLMs) are widely used to tackle complex tasks with autonomous workflows. Recently, reusable natural language skills have emerged as a popular paradigm to inject procedural knowledge into LLM applications. Since popular skills are often invoked repeatedly, placing their full text in every context significantly increases prefill cost and latency. While text compression techniques have the potential to solve this problem, most existing methods are designed to compress factual knowledge in documents instead of procedural knowledge, making them insufficient for skill compression. In this paper, we argue that an effective skill compression method should: 1) preserve logical dependencies among workflows and tool protocols, 2) enable lightweight, offline compression for frequently updated community skills, and 3) be adaptable to varying complexities across skills. To address this, we present SKIM (SKIll coMpression), an adaptive multi-resolution soft token compression framework for procedural skills. Depending on the complexity of each skill, SKIM creates different numbers of soft tokens that not only improve the efficiency of LLM inference, but also preserve the effectiveness of skill usage. Experiments indicate that SKIM compresses skills to 30 to 60 percent of their original token length while preserving task performance better than existing compression methods.We have released our code at https://github.com/bebr2/SKIM .
Community
Can LLM skills be compressed without losing the procedural knowledge that makes them executable? SKIM adaptively represents each skill with multi-resolution soft tokens, preserving workflows, logical dependencies, and tool-use protocols while reducing context usage to 30–60%. This provides an initial step toward more compact and reusable representations of procedural knowledge for LLM agents.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2606.12203 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.12203 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.12203 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.