© 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. 63521
    Prev
    Next
    INDEX
    Explanations

    letters and symbols that represent specific identifiers or categories

    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
    adays
    -0.27
    odore
    -0.22
    xiety
    -0.21
    bidden
    -0.21
    bsites
    -0.20
    anmar
    -0.19
    greSQL
    -0.18
    esda
    -0.18
    wealth
    -0.18
    gether
    -0.18
    POSITIVE LOGITS
    orem
    0.16
    sWith
    0.14
    /OR
    0.14
    noDB
    0.13
    vas
    0.13
    /Object
    0.13
     Mane
    0.13
    /Instruction
    0.13
    SetText
    0.13
    #echo
    0.12
    Activations Density 0.002%

    No Known Activations