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. 23179
    Prev
    Next
    INDEX
    Explanations

    phrases and terms related to dosage and preparation in the context of substances

    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
    amedi
    -0.14
    (s
    -0.13
    ноп
    -0.13
     dint
    -0.13
     Millions
    -0.13
     billions
    -0.13
     cigaret
    -0.13
    argin
    -0.13
    /by
    -0.12
     vat
    -0.12
    POSITIVE LOGITS
     generally
    0.26
     anything
    0.25
     Anything
    0.24
    Anything
    0.24
     typical
    0.23
     larger
    0.23
     higher
    0.23
     Generally
    0.23
     aim
    0.21
    Typ
    0.21
    Activations Density 0.110%

    No Known Activations