© 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. 20-TRANSCODER-HP
    7. 62772
    Prev
    Next
    INDEX
    Explanations

    Technology, software, economics

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_20.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.20.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Negative Logits
    ONEY
    -0.26
    tails
    -0.26
    enko
    -0.26
    nock
    -0.25
    ÑijÑĢ
    -0.25
    (digits
    -0.24
    黥
    -0.24
    -vars
    -0.24
    vak
    -0.24
    ANTS
    -0.24
    POSITIVE LOGITS
    ä½Ĩå®ŀéĻħä¸Ĭ
    0.35
    âĢ¦but
    0.33
     indeed
    0.32
    亿åħĥ以ä¸Ĭ
    0.32
     cor
    0.32
    ä¸įä»ħä»ħæĺ¯
    0.31
    —but
    0.31
     but
    0.31
     actually
    0.30
    ä¸įä¸Ģå®ļ
    0.30
    Activations Density 31.820%

    No Known Activations