© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 28965
    Prev
    Next
    INDEX
    Explanations

    mentions of the character Luke Skywalker

    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
     tract
    -0.17
    iner
    -0.15
    line
    -0.15
    ride
    -0.15
    refund
    -0.14
    Äįet
    -0.14
    isse
    -0.14
    RYPTO
    -0.14
    ounge
    -0.14
    éĩı
    -0.14
    POSITIVE LOGITS
     Skywalker
    0.38
     Cage
    0.20
    ovsky
    0.18
    irst
    0.18
    hiro
    0.16
    Sky
    0.15
     Luke
    0.15
     sky
    0.15
    icrous
    0.15
    athing
    0.15
    Activations Density 0.003%

    No Known Activations