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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & 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. 43
    Prev
    Next
    INDEX
    Explanations

    punctuation marks, particularly question marks and quotation marks, indicating dialogue or queries

    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
    اتÛĮ
    -0.16
    ãģįãģŁ
    -0.15
    -caret
    -0.14
    ich
    -0.14
    niest
    -0.13
    ings
    -0.13
    nels
    -0.13
    -scrollbar
    -0.13
     omas
    -0.13
    niej
    -0.13
    POSITIVE LOGITS
    ed
    0.23
     "
    0.19
    i
    0.18
    ÂĿ
    0.16
    ing
    0.16
    aft
    0.15
    Ø©
    0.15
    651
    0.15
    -than
    0.15
    al
    0.15
    Activations Density 0.054%

    No Known Activations