© 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. 30-TRANSCODER-HP
    7. 95138
    Prev
    Next
    INDEX
    Explanations

    hand

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    YTE
    -0.28
    Bring
    -0.26
    LOPT
    -0.26
     leveling
    -0.26
    utow
    -0.26
    bring
    -0.26
    yun
    -0.25
    修补
    -0.25
    .Parcelable
    -0.24
    /general
    -0.24
    POSITIVE LOGITS
     held
    0.35
     hold
    0.33
    Hold
    0.33
    åıij
    0.32
     Hold
    0.32
    hold
    0.31
     holds
    0.29
     holding
    0.29
     behind
    0.29
     Positioned
    0.28
    Activations Density 0.015%

    No Known Activations