© 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. 24691
    Prev
    Next
    INDEX
    Explanations

    instances of words that signify change or newness

    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
    Ñĥли
    -0.16
    orelease
    -0.14
    :black
    -0.14
    .ribbon
    -0.14
    unifu
    -0.13
    borg
    -0.13
    δά
    -0.13
    udur
    -0.13
    intree
    -0.13
    .habbo
    -0.12
    POSITIVE LOGITS
     light
    1.60
     Light
    1.34
    light
    1.28
    Light
    1.19
     LIGHT
    1.14
    -light
    1.14
    _light
    1.02
    LIGHT
    1.00
     lights
    0.94
    .light
    0.93
    Activations Density 0.014%

    No Known Activations