© 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. 43880
    Prev
    Next
    INDEX
    Explanations

    phrases that convey a sense of elegance and beauty, particularly in relation to music

    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
     thin
    -0.16
     unique
    -0.15
     awkward
    -0.15
     isolated
    -0.15
     bold
    -0.15
     otherwise
    -0.15
    @brief
    -0.15
     Thin
    -0.14
     simple
    -0.14
     weak
    -0.14
    POSITIVE LOGITS
    racak
    0.16
    ERO
    0.15
     echt
    0.15
    atego
    0.15
     fours
    0.15
    chwitz
    0.15
    .IContainer
    0.15
     Deploy
    0.14
    PLOY
    0.14
    .appendTo
    0.14
    Activations Density 0.012%

    No Known Activations