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. 23134
    Prev
    Next
    INDEX
    Explanations

    references to hierarchical relationships and rankings

    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
    likes
    -0.15
    овиÑĩ
    -0.14
    åĿª
    -0.14
    ovich
    -0.14
     بخشÛĮ
    -0.14
    äºĪ
    -0.14
    remainder
    -0.13
     likes
    -0.13
    à¸Ħวามร
    -0.13
     entirety
    -0.13
    POSITIVE LOGITS
     many
    0.36
     few
    0.31
    many
    0.28
     several
    0.27
     most
    0.26
    åĩłä¸ª
    0.23
     Many
    0.22
    few
    0.22
    Many
    0.22
     MANY
    0.22
    Activations Density 0.102%

    No Known Activations