© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 33406
    Prev
    Next
    INDEX
    Explanations

    instructions and recommendations related to hair care

    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
     cigaret
    -0.14
     ún
    -0.14
    Pie
    -0.14
    engan
    -0.13
     pie
    -0.13
     welded
    -0.13
    amus
    -0.13
    íݸ
    -0.13
    amient
    -0.13
    ulta
    -0.13
    POSITIVE LOGITS
     mix
    0.32
     Mix
    0.29
    mix
    0.28
     apply
    0.26
     mixed
    0.26
    Mix
    0.25
     pour
    0.25
     mixes
    0.23
     mixing
    0.23
     use
    0.22
    Activations Density 0.118%

    No Known Activations