© 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. 54772
    Prev
    Next
    INDEX
    Explanations

    code and licenses

    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
    ersh
    -0.31
    åĬ¡
    -0.28
     topo
    -0.28
    åĭĻ
    -0.27
    pek
    -0.26
    å®ĥçļĦ
    -0.25
    riel
    -0.25
     matters
    -0.25
    ria
    -0.25
    _mob
    -0.24
    POSITIVE LOGITS
     norske
    0.26
    nonnull
    0.25
     ton
    0.24
    궤
    0.24
    èݪ
    0.24
     leaving
    0.24
    ä¸Ģèµ·æĿ¥
    0.23
     tank
    0.23
    éĤķ
    0.23
    .palette
    0.23
    Activations Density 0.016%

    No Known Activations