gary hsu
    • Create new note
    • Create a note from template
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Write
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note No publishing access yet

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.

      Your account was recently created. Publishing will be available soon, allowing you to share notes on your public page and in search results.

      Your team account was recently created. Publishing will be available soon, allowing you to share notes on your public page and in search results.

      Explore these features while you wait
      Complete general settings
      Bookmark and like published notes
      Write a few more notes
      Complete general settings
      Write a few more notes
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights New
    • Engagement control
    • Make a copy
    • Transfer ownership
    • Delete this note
    • Save as template
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Note Insights Versions and GitHub Sync Sharing URL Create Help
Create Create new note Create a note from template
Menu
Options
Engagement control Make a copy Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Write
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note No publishing access yet

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.

    Your account was recently created. Publishing will be available soon, allowing you to share notes on your public page and in search results.

    Your team account was recently created. Publishing will be available soon, allowing you to share notes on your public page and in search results.

    Explore these features while you wait
    Complete general settings
    Bookmark and like published notes
    Write a few more notes
    Complete general settings
    Write a few more notes
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       Owned this note    Owned this note      
    Published Linked with GitHub
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    # [MIT 新研究] - 遞迴語言模型 (Recursive Language Model, RLM):突破 100 萬 Token 的推理新典範 本筆記探討了由 MIT 團隊提出的最新技術:**遞迴語言模型 (Recursive Language Model, RLM)**。這是一種全新的推理方法論(Inference Methodology),旨在解決大型語言模型(LLM)在處理長文本時常見的「上下文腐敗」(Context Rot)與推理能力下降問題,讓如 **GPT-5** 等模型能實現在 100 萬甚至 1,000 萬 Token 級別下的精準推理。 --- ## 1. ⚠️ 現有長文本模型的困境:上下文腐敗 (Context Rot) * **性能衰減**:隨著上下文長度增加,LLM 的推理能力並非線性下降,而是劇烈衰退,特別是在科學、金融與醫學等資訊密集型任務中。 * **大海撈針 (Needle in a Haystack) 的誤導**:現有的 100% 準確率測試僅代表「檢索」能力(複雜度為 1),而非真正的「推理」能力。 * **推理極限**:研究顯示 **GPT-5** 在處理具有線性複雜度(O(n))或平方複雜度(O(n²))的推理任務時,性能在 16k Token 處開始下滑,到 33k Token 時幾乎歸零。 * **相變現象 (Phase Transition)**:當資訊密度增加,傳統 Transformer 的注意力機制(Attention Mechanism)會失去連貫性,無法有效處理複雜邏輯。 > **核心定義:上下文腐敗 (Context Rot)** > 指的是 LLM 雖然擁有巨大的上下文窗口(如 272k),但在處理長文本時會產生幻覺或失去焦點,導致其在長文本中的推理品質遠低於短文本。 --- ## 2. 🌀 遞迴語言模型 (RLM):神經符號系統的解決方案 MIT 的核心概念是將長 Prompt 視為「外部環境」,而非直接餵入神經網路。 * **外部環境化**:將長達數十萬字元的 Prompt 視為硬碟上的文件,LLM 則扮演 CPU。 * **神經符號交互 (Neuro-symbolic Interaction)**:LLM 不直接處理原始數據,而是寫出 **Python** 程式碼來對文本進行切片、索引與操作。 * **虛擬記憶體機制**:模仿作業系統,將數據分頁(Paging)調入記憶體處理,處理完後清空,再加載下一頁,從而繞過 Transformer 的 Token 限制。 --- ## 3. 🛠️ RLM 的結構創新與執行階段 RLM 的運作依賴於三個關鍵組件:環境(Python REPL)、介面(LLM Query 函數)與遞迴軌跡。 ### 執行的四個階段: 1. **探測 (Probing)**:LLM 編寫 Python 程式碼(如 **Regex** 或切片),檢查數據結構(例如:先看 CSV 的前五行)。 2. **分解 (Decomposition)**:LLM 寫出迴圈(Loop),決定如何遍歷定義好的上下文。 3. **遞迴 (Recursion)**:在迴圈中調用 `LLM_query` 函數,生成一個「乾淨」的 LLM 子實例來處理特定片段。 4. **聚合 (Aggregation)**:根模型(Root LLM)收集所有子實例的結果,合成最終答案。 > **結構亮點:確定性 (Deterministic)** > 傳統 **RAG** 是機率性的(依賴向量檢索的相似度);而 **RLM** 是確定性的,它透過程式碼遍歷 100% 的數據,確保無遺漏。 --- ## 4. 📊 性能表現與成本分析 RLM 在複雜任務上展示了驚人的進步,尤其是在處理 **Olong Pairs** (尋找所有特徵配對) 等高難度基準測試時。 * **性能對比**: * 在 32k Token 的平方複雜度任務中,原版 **GPT-5** 成功率僅 **0.04%**。 * 使用 RLM 封裝後的 **GPT-5** 成功率飆升至 **58%**。 * **成本效益**: * 雖然 API 調用成本會因遞迴與多次採樣而增加(例如從 16 美分增加到 33 美分),但換取的是從「完全不可用」到「高度可用」的飛躍。 * 即使上下文達到 100 萬 Token,非推理型任務的成本仍保持穩定,僅高難度推理任務成本會隨複雜度上升。 --- ## 5. 💡 2026 AI 發展新趨勢:神經符號化 這項研究標誌著 Transformer 的「注意力機制」並非萬能。 * **角色轉變**: * **LLM (神經部分)**:負責模糊直覺、語義理解、生成程式碼與撰寫總結。 * **程式碼 (符號部分)**:負責剛性邏輯、計數、迴圈與精確索引。 * **未來展望**:與其追求更大的預訓練窗口,不如透過 **非同步並行 (Async Parallelism)** 最佳化遞迴循環的延遲。 > **專家評論** > 「神經網路目前不足以支撐大規模的精確推理,必須依賴『符號外骨骼 (Symbolic Exoskeleton)』的支撐。RLM 實質上是將記憶問題轉化為了管理問題。」 ---

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password
    or
    Sign in via Facebook Sign in via X(Twitter) Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    By signing in, you agree to our terms of service.

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully