© 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. 30488
    Prev
    Next
    INDEX
    Explanations

    the repetition of the word "and" and similar conjunctions

    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
    luet
    -0.14
    /***/
    -0.13
     “âĢ¦
    -0.13
    ,↵↵↵↵
    -0.13
    âĢª
    -0.13
    ÑģилÑĮ
    -0.12
     ?><?
    -0.12
    ï¸
    -0.12
     deÄŁil
    -0.12
    ÑģÑĮ
    -0.12
    POSITIVE LOGITS
     even
    0.38
    /or
    0.37
     other
    0.33
     etc
    0.30
     finally
    0.30
    even
    0.28
     hatta
    0.27
     downright
    0.26
    rog
    0.25
    etc
    0.25
    Activations Density 0.436%

    No Known Activations