© 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. 96910
    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_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
    èĭĵ
    -0.28
    çľ¼éĥ¨
    -0.25
    rovers
    -0.24
    çļĦ第ä¸Ģ个
    -0.24
    ìĿij
    -0.24
    -wall
    -0.24
    appers
    -0.24
    æľĢå°ij
    -0.23
    kp
    -0.23
    æĪ¡
    -0.23
    POSITIVE LOGITS
     Nation
    0.29
    两
    0.28
     nation
    0.28
    (low
    0.27
     two
    0.27
     lead
    0.26
    oton
    0.26
    LOW
    0.26
    unders
    0.25
     Recall
    0.25
    Activations Density 0.108%

    No Known Activations