© 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. 25979
    Prev
    Next
    INDEX
    Explanations

    occurrences of legal terminology and references to moral or ethical issues in the context of state actions

    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
     SUBSTITUTE
    -0.12
    ¶Į
    -0.12
    _$_
    -0.12
     torino
    -0.11
     dostan
    -0.11
    "class
    -0.11
    кид
    -0.11
     seins
    -0.11
     âĶľ
    -0.11
     ÑĩиÑģлÑĸ
    -0.11
    POSITIVE LOGITS
     see
    0.18
     (
    0.16
     reference
    0.16
     thus
    0.16
    references
    0.16
     b
    0.15
     therefore
    0.15
     further
    0.15
     these
    0.14
    References
    0.14
    Activations Density 0.774%

    No Known Activations