© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Qwen3-4B
    5. 20-TRANSCODER-HP
    7. 129723
    Prev
    Next
    INDEX
    Explanations

    programming

    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
    Not in Any Lists

    No Comments

    Negative Logits
    otten
    -0.30
    acos
    -0.25
    ament
    -0.25
    abus
    -0.24
     turning
    -0.24
     standing
    -0.24
    ilmiÅŁ
    -0.24
    å¼Ģå±Ģ
    -0.24
    arring
    -0.23
    çī¹æľĹ
    -0.23
    POSITIVE LOGITS
     cig
    0.26
    esti
    0.26
    ests
    0.26
    åĮº
    0.26
    éĢĶ
    0.25
    zioni
    0.25
    theses
    0.25
    thumbs
    0.25
    borg
    0.25
    声
    0.24
    Activations Density 27.826%

    No Known Activations