© 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. 129890
    Prev
    Next
    INDEX
    Explanations
    No Explanations Found
    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
    æĪ¿å±ĭ
    -0.26
     trò
    -0.25
    ''↵
    -0.24
     dames
    -0.23
    [âĢ¦
    -0.23
    olute
    -0.23
     MyClass
    -0.23
    mj
    -0.23
    maj
    -0.23
    åľ°çľĭçĿĢ
    -0.22
    POSITIVE LOGITS
    éħįåIJĪ
    0.26
    è¨Ģ
    0.25
    owering
    0.24
    举
    0.24
    ÑĢг
    0.23
    æģ¯
    0.23
    äºĴ
    0.23
     NF
    0.23
     RID
    0.22
    ony
    0.22
    Activations Density 13.730%

    No Known Activations

    This feature has no known activations.