© 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. 141873
    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
    ocal
    -0.27
    aveled
    -0.27
    åĿ¨
    -0.26
    ç½IJ
    -0.25
    enticator
    -0.24
    çģ«
    -0.24
    qualified
    -0.24
    orts
    -0.23
    ismic
    -0.23
    ritt
    -0.23
    POSITIVE LOGITS
    åºĩ
    0.28
    éĺ³åİ¿
    0.28
    æĶ¶æķĽ
    0.27
    çķĢ
    0.26
    zos
    0.25
     uda
    0.25
    ä¼łåĩº
    0.24
    yah
    0.24
    hid
    0.24
     Rak
    0.24
    Activations Density 8.081%

    No Known Activations

    This feature has no known activations.