© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 24706
    Prev
    Next
    INDEX
    Explanations

    punctuation marks and their surrounding context

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    çĵľ
    -0.15
    Js
    -0.14
     onCancelled
    -0.13
    漫çĶ»
    -0.13
    pte
    -0.13
     Giz
    -0.12
    imoto
    -0.12
    ropoda
    -0.12
    Emoji
    -0.12
     smarty
    -0.12
    POSITIVE LOGITS
     game
    0.42
     games
    0.35
    game
    0.31
     Game
    0.30
    -game
    0.30
     gameplay
    0.28
    Game
    0.27
    游æĪı
    0.25
     GAME
    0.25
     Games
    0.25
    Activations Density 0.005%

    No Known Activations