Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23357
    Prev
    Next
    INDEX
    Explanations

    expressions of knowledge or awareness

    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
    ån
    -0.16
    umer
    -0.15
    ÑĢÑĸд
    -0.15
    andalone
    -0.13
    forms
    -0.13
    iten
    -0.13
    .modelo
    -0.13
    usto
    -0.13
    aná
    -0.13
    ebi
    -0.13
    POSITIVE LOGITS
     perfectly
    0.36
     well
    0.29
     exactly
    0.29
     that
    0.27
     full
    0.27
     better
    0.26
     precisely
    0.25
     how
    0.23
     what
    0.21
     too
    0.21
    Activations Density 0.072%

    No Known Activations