© 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. 12704
    Prev
    Next
    INDEX
    Explanations

    expressions of emotional intensity or significance

    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
    oram
    -0.13
    erais
    -0.13
    ANGO
    -0.13
    mad
    -0.13
    øre
    -0.13
    agina
    -0.13
    ilon
    -0.13
    ialis
    -0.13
     incom
    -0.13
    ckill
    -0.12
    POSITIVE LOGITS
     fitting
    0.81
     appropriate
    0.69
    appropriate
    0.61
    fit
    0.59
     apt
    0.56
     appropri
    0.53
     Appropri
    0.46
     fits
    0.44
    .fit
    0.43
     suitable
    0.43
    Activations Density 0.032%

    No Known Activations