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. 23598
    Prev
    Next
    INDEX
    Explanations

    the presence of phrases emphasizing relationships or connections

    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
    buf
    -0.15
    woord
    -0.14
    enci
    -0.14
    anoi
    -0.14
    511
    -0.14
    .Emit
    -0.13
    Ñıж
    -0.13
    gaard
    -0.13
    oland
    -0.13
    Bu
    -0.13
    POSITIVE LOGITS
     bring
    1.24
     Bring
    1.10
     brings
    1.10
     bringing
    1.10
    bring
    1.10
     brought
    1.09
    Bring
    1.07
    bringing
    0.99
     Bringing
    0.82
    rought
    0.72
    Activations Density 0.006%

    No Known Activations