© 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. 22020
    Prev
    Next
    INDEX
    Explanations

    concepts related to banking restrictions and access to ATMs

    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
    ãĤ·ãĤ§
    -0.15
    lices
    -0.15
    precated
    -0.15
    radan
    -0.14
     Watts
    -0.14
    stral
    -0.14
    ãĥ¼ãĥ©
    -0.14
    æĿ¯
    -0.14
    izzas
    -0.14
    ouri
    -0.14
    POSITIVE LOGITS
     ATM
    0.47
     AT
    0.41
    AT
    0.40
     withdraw
    0.34
     withdrawals
    0.34
     withdrawal
    0.33
     atm
    0.33
     branch
    0.33
    _AT
    0.32
     Withdraw
    0.32
    Activations Density 0.021%

    No Known Activations