© 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. 31-TRANSCODER-HP
    7. 18227
    Prev
    Next
    INDEX
    Explanations

    CDC

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     straw
    -0.30
    ä¸įæĪIJ
    -0.28
    ll
    -0.26
    头æĿ¡
    -0.26
    å¼Ģåı£
    -0.26
     claw
    -0.26
    çłģ
    -0.25
    樽
    -0.24
    ç´¢èµĶ
    -0.24
     angle
    -0.24
    POSITIVE LOGITS
     Mat
    0.31
    ision
    0.28
     Guidance
    0.28
    Mat
    0.27
    iosa
    0.27
    -tracking
    0.27
    ISION
    0.26
    MM
    0.26
     Guid
    0.25
    iative
    0.25
    Activations Density 0.037%

    No Known Activations