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. 23252
    Prev
    Next
    INDEX
    Explanations

    references to the reader or audience

    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
     Souls
    -0.16
    inus
    -0.15
    tat
    -0.15
    agic
    -0.14
    lover
    -0.14
    tember
    -0.14
    esti
    -0.14
    że
    -0.14
    ÏĦÎŃ
    -0.14
    ãĥĭãĥĥãĤ¯
    -0.14
    POSITIVE LOGITS
     heard
    0.23
     know
    0.21
     bet
    0.21
     mileage
    0.19
     knew
    0.17
    Bet
    0.16
     can
    0.16
    heard
    0.16
     Bet
    0.16
     better
    0.16
    Activations Density 0.048%

    No Known Activations