© 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. 45111
    Prev
    Next
    INDEX
    Explanations

    actions related to taking responsibility or initiative

    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
    PU
    -0.17
     perse
    -0.14
    ãģĤãģĴ
    -0.14
    iva
    -0.14
    entiful
    -0.14
    uja
    -0.14
    å¸Ī
    -0.13
    elp
    -0.13
    .easing
    -0.13
    /job
    -0.13
    POSITIVE LOGITS
     responsibility
    0.31
     ownership
    0.28
     charge
    0.28
     stock
    0.27
     advantage
    0.27
     steps
    0.25
     Responsibility
    0.24
     inventory
    0.24
    Ownership
    0.24
    ownership
    0.23
    Activations Density 0.055%

    No Known Activations