© 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. 108311
    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
    çľĭä¼¼
    -0.27
    ined
    -0.27
    ignet
    -0.27
    带éĺŁ
    -0.25
     Bot
    -0.25
     grav
    -0.25
    éd
    -0.25
    umper
    -0.24
    edly
    -0.24
     Morton
    -0.24
    POSITIVE LOGITS
    istik
    0.26
    æĽ²
    0.25
    COMMENT
    0.24
    åħĥ
    0.23
     Conditioning
    0.23
    çŀŃ解
    0.23
    idad
    0.23
    äºĮæīĭ
    0.22
    PTR
    0.22
     Courier
    0.22
    Activations Density 43.802%

    No Known Activations

    This feature has no known activations.