© 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. 35-TRANSCODER-HP
    7. 39480
    Prev
    Next
    INDEX
    Explanations

    say "ad"

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_35.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.35.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    仪
    -0.28
     roman
    -0.24
     bows
    -0.24
     goof
    -0.24
     bowling
    -0.24
     Beng
    -0.24
    æĭ¥æĮ¤
    -0.23
     bow
    -0.23
    imest
    -0.23
    樱æ¡ĥ
    -0.23
    POSITIVE LOGITS
    ads
    0.82
    ad
    0.81
    aded
    0.80
    ades
    0.80
    ade
    0.77
    ading
    0.77
    adas
    0.77
    aden
    0.75
    adx
    0.74
    ader
    0.73
    Activations Density 74.710%

    No Known Activations