© 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. 44770
    Prev
    Next
    INDEX
    Explanations

    references to programming concepts related to escaping characters

    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
     Hos
    -0.14
     decentral
    -0.14
     ("\
    -0.14
    lic
    -0.14
    ussy
    -0.13
    ves
    -0.13
     publicly
    -0.13
     |[
    -0.13
    ught
    -0.13
    le
    -0.13
    POSITIVE LOGITS
    ardu
    0.15
    ilinx
    0.14
    ept
    0.14
    rire
    0.14
    orney
    0.14
     AssemblyVersion
    0.13
    estion
    0.13
    .hxx
    0.13
    utsch
    0.13
    ancial
    0.13
    Activations Density 0.104%

    No Known Activations