© 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. 32-TRANSCODER-HP
    7. 44213
    Prev
    Next
    INDEX
    Explanations

    physical

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_32.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.32.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    å½ķ
    -0.33
    uten
    -0.29
    ytt
    -0.26
    chrift
    -0.26
    (mut
    -0.26
    éĩĮæĸ¯
    -0.25
    积
    -0.24
    oct
    -0.24
    éĵº
    -0.24
    åį·
    -0.24
    POSITIVE LOGITS
     exam
    0.30
    alsex
    0.29
     nghiá»ĩp
    0.29
     SEX
    0.28
     Splash
    0.28
    sthrough
    0.28
     exams
    0.27
    allel
    0.26
    sex
    0.26
    SEX
    0.26
    Activations Density 0.002%

    No Known Activations