© 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. Qwen3-4B
    5. 22-TRANSCODER-HP
    7. 112093
    Prev
    Next
    INDEX
    Explanations

    punctuation

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_22.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.22.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Desk
    -0.26
    bum
    -0.25
    ertools
    -0.25
    ä¾µ
    -0.24
    -F
    -0.24
    азвание
    -0.24
     Compet
    -0.24
    ussy
    -0.23
    kü
    -0.23
    å°Ĩèĩªå·±çļĦ
    -0.23
    POSITIVE LOGITS
    åĴĮåıijå±ķ
    0.26
    кал
    0.24
    ä»ĸ们çļĦ
    0.24
    antly
    0.24
    wb
    0.24
     uda
    0.23
    åĴĮæĬĢæľ¯
    0.23
     Walnut
    0.23
    their
    0.23
    Others
    0.23
    Activations Density 0.006%

    No Known Activations