© 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. 64349
    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
    é«ĺåľ°
    -0.26
    ,set
    -0.25
    罩
    -0.25
     NOTHING
    -0.24
    .loading
    -0.24
     Modi
    -0.23
     setSelected
    -0.23
    _PARAMS
    -0.23
     MIS
    -0.23
    shal
    -0.23
    POSITIVE LOGITS
    ä¸įä¸ĭ
    0.30
    tte
    0.28
    issement
    0.25
    mate
    0.25
    ophone
    0.25
    eat
    0.25
     ella
    0.25
    à¸Ńà¸Ķ
    0.24
    pper
    0.24
    eve
    0.24
    Activations Density 1.363%

    No Known Activations

    This feature has no known activations.