© 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. 63338
    Prev
    Next
    INDEX
    Explanations

    with

    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
     neutrality
    -0.30
    è¿ĻéĩĮæľī
    -0.28
    yı
    -0.26
     nâ
    -0.26
    è¿ĻéĩĮæĺ¯
    -0.26
     Sinclair
    -0.25
    ihan
    -0.25
    neutral
    -0.25
    atts
    -0.24
     Neutral
    -0.24
    POSITIVE LOGITS
    MODULE
    0.27
    èķĥ
    0.27
    IOR
    0.27
     undergoing
    0.27
     pap
    0.27
    icle
    0.26
     offending
    0.26
    åı£æ°´
    0.25
    orris
    0.25
    amos
    0.24
    Activations Density 3.674%

    No Known Activations