© 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. 33-TRANSCODER-HP
    7. 54577
    Prev
    Next
    INDEX
    Explanations

    the

    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
    Negative Logits
    upper
    -0.27
    iasm
    -0.26
    adge
    -0.26
     desk
    -0.26
    adow
    -0.26
     Adler
    -0.25
     Clo
    -0.25
    ,proto
    -0.25
    è¶ĭ
    -0.24
    desk
    -0.24
    POSITIVE LOGITS
    WN
    0.26
    ç½
    0.26
    OME
    0.24
     Sn
    0.24
     scratch
    0.24
    chan
    0.24
    UTILITY
    0.24
    åķľ
    0.23
    _pars
    0.23
    (actual
    0.23
    Activations Density 0.025%

    No Known Activations