© 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. 130001
    Prev
    Next
    INDEX
    Explanations

    statistics and numbers

    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
    FK
    -0.28
    acha
    -0.27
    ãĢģ“
    -0.23
     surg
    -0.23
     oo
    -0.23
    SOC
    -0.23
    羣çļ®
    -0.23
     Spatial
    -0.23
    wife
    -0.23
    èĩª
    -0.23
    POSITIVE LOGITS
    客æĪ·æıIJä¾Ľ
    0.26
    åħ¨åľº
    0.25
     ).↵↵
    0.25
    wers
    0.25
    thin
    0.25
    为æĤ¨æıIJä¾Ľ
    0.24
     complete
    0.24
    æĭľ
    0.24
    authorized
    0.24
    鸦
    0.24
    Activations Density 13.930%

    No Known Activations