© 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. 108114
    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
    ewish
    -0.26
    /U
    -0.26
    幸
    -0.25
    ç»Ļä»ĸ们
    -0.25
    ble
    -0.25
    峪
    -0.25
    	U
    -0.25
    gue
    -0.24
    toHaveLength
    -0.24
    .U
    -0.23
    POSITIVE LOGITS
    ä¸ĭåįĬ
    0.29
    åIJĮçŃī
    0.28
    å¸ĺ
    0.27
     bele
    0.26
    uyên
    0.25
    -sem
    0.25
    ħ§
    0.24
     Denied
    0.24
    ä¸īçĤ¹
    0.24
    atat
    0.24
    Activations Density 57.846%

    No Known Activations

    This feature has no known activations.