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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 8466
    Prev
    Next
    INDEX
    Explanations

    prepositions and locational phrases

    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
    çļĦä¸Ģ个
    -0.26
    æĺ¯ä¸Ģ个
    -0.24
    ä¸Ģ个
    -0.23
    ä¸ĢåĢĭ
    -0.20
     sebuah
    -0.20
     another
    -0.19
     eines
    -0.16
     einer
    -0.16
    another
    -0.15
     ενÏĮÏĤ
    -0.15
    POSITIVE LOGITS
     different
    0.28
     various
    0.28
     other
    0.27
     which
    0.24
     Various
    0.22
    ä¸įåIJĮ
    0.21
     both
    0.20
    ä¸įåIJĮçļĦ
    0.20
     коÑĤоÑĢÑĭе
    0.20
    different
    0.20
    Activations Density 0.079%

    No Known Activations