© 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. 33-TRANSCODER-HP
    7. 53170
    Prev
    Next
    INDEX
    Explanations

    github

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_33.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.33.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    æĿľ
    -0.31
    olta
    -0.25
     migrated
    -0.25
    å¼ĤæĢ§
    -0.25
    mph
    -0.24
    éĺ³åİ¿
    -0.24
    宪
    -0.24
    ứng
    -0.24
     em
    -0.24
    ocumented
    -0.24
    POSITIVE LOGITS
    æĮĤ
    0.26
    ATUS
    0.26
     tours
    0.26
     Quotes
    0.25
    èĢģåŃIJ
    0.24
    æĭįåįĸ
    0.24
    _tC
    0.23
    工夫
    0.23
    ctic
    0.23
    NIC
    0.23
    Activations Density 0.002%

    No Known Activations