© 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. 24899
    Prev
    Next
    INDEX
    Explanations

    specific actions or processes related to roles, activities, and events

    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
     ÑĸнÑĪими
    -0.15
     дÑĢÑĥгиÑħ
    -0.15
     дÑĢÑĥгими
    -0.15
     other
    -0.14
    ä¸ĢåĪĩ
    -0.14
    опол
    -0.14
    าล
    -0.14
    åħ¶ä»ĸ
    -0.14
    ì²Ļ
    -0.14
     dalÅ¡ÃŃch
    -0.14
    POSITIVE LOGITS
     a
    0.45
    ä¸Ģ个
    0.40
     sebuah
    0.39
     an
    0.39
     eine
    0.33
    ä¸ĢåĢĭ
    0.32
     een
    0.32
    —a
    0.28
    æĺ¯ä¸Ģ个
    0.28
    çļĦä¸Ģ个
    0.26
    Activations Density 0.682%

    No Known Activations