© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 32-TRANSCODER-HP
    7. 41013
    Prev
    Next
    INDEX
    Explanations

    tools

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    quat
    -0.27
    éĿĪ
    -0.27
    .notify
    -0.26
    çģµ
    -0.26
    ibrary
    -0.26
     lex
    -0.25
     lanc
    -0.25
    ä¸ĢåIJ¬
    -0.25
     building
    -0.25
    奥æŀĹ
    -0.25
    POSITIVE LOGITS
    Envelope
    0.28
    uchi
    0.27
    piel
    0.26
     Vest
    0.25
     sz
    0.25
    æľĢåIJİä¸Ģ次
    0.24
    Sz
    0.24
    conds
    0.24
    @s
    0.24
     coisa
    0.24
    Activations Density 0.003%

    No Known Activations