© 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. 54692
    Prev
    Next
    INDEX
    Explanations

    Code and URLs

    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
    uations
    -0.30
    composition
    -0.28
    utations
    -0.27
    gings
    -0.27
    olation
    -0.26
     subtraction
    -0.26
     commenced
    -0.26
    subtract
    -0.26
    Descriptions
    -0.26
    Purpose
    -0.26
    POSITIVE LOGITS
     it
    0.35
     a
    0.33
    er
    0.29
    å¼ı
    0.28
    an
    0.28
     away
    0.28
    å®ĥ
    0.27
     he
    0.27
     re
    0.26
    od
    0.26
    Activations Density 0.087%

    No Known Activations