© 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. 20-TRANSCODER-HP
    7. 129788
    Prev
    Next
    INDEX
    Explanations

    Negative or undesirable situations

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ponsible
    -0.29
    -SA
    -0.26
     BaseService
    -0.26
    èĥ½æ»¡è¶³
    -0.26
    æĽ´å¥½çļĦ
    -0.25
    ibrator
    -0.25
     Ment
    -0.25
    çļĦæķ´ä½ĵ
    -0.24
    Piece
    -0.24
    urovision
    -0.23
    POSITIVE LOGITS
    ä¼Łå¤§
    0.28
    ople
    0.26
    æĹ¶ä»£
    0.26
    ÑĢад
    0.25
    ä¸¥æł¼
    0.25
     maj
    0.25
    rim
    0.25
    alf
    0.24
     cancell
    0.24
    umbing
    0.24
    Activations Density 25.639%

    No Known Activations