© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 44482
    Prev
    Next
    INDEX
    Explanations

    key events and statistics in sports

    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
    rahim
    -0.17
    aliz
    -0.15
    zyst
    -0.14
    Ìģc
    -0.14
     DefaultValue
    -0.14
    ắp
    -0.14
    _nick
    -0.14
    ayah
    -0.14
    -LAST
    -0.14
    .btnSave
    -0.14
    POSITIVE LOGITS
     courtesy
    0.19
     unanswered
    0.18
    anson
    0.17
     Early
    0.16
     early
    0.16
     Courtesy
    0.16
     via
    0.15
     followed
    0.15
     aprove
    0.15
     tall
    0.15
    Activations Density 0.027%

    No Known Activations