© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 21253
    Prev
    Next
    INDEX
    Explanations

    phrases indicating negative or harmful actions and outcomes

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Negative Logits
    $MESS
    -0.18
    $__
    -0.16
    $LANG
    -0.16
     dikke
    -0.16
    .googleapis
    -0.14
    trak
    -0.14
     nackte
    -0.14
    ATTERN
    -0.13
    ERCHANT
    -0.13
    gend
    -0.13
    POSITIVE LOGITS
    (sp
    0.21
    0.17
    esda
    0.13
    oned
    0.13
    gether
    0.13
    ÑĥÑĢе
    0.13
    еÑĢÑĮ
    0.12
    ÑģоÑĢ
    0.12
     forth
    0.12
     ArgumentOutOfRangeException
    0.12
    Activations Density 0.508%

    No Known Activations