© 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. 30-TRANSCODER-HP
    7. 96009
    Prev
    Next
    INDEX
    Explanations

    comma

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_30.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.30.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     which
    -0.61
    which
    -0.57
     Which
    -0.52
     коÑĤоÑĢого
    -0.50
     коÑĤоÑĢом
    -0.49
    Which
    -0.48
    åĵªç§į
    -0.47
    .which
    -0.47
    åĵª
    -0.47
     WHICH
    -0.43
    POSITIVE LOGITS
     who
    0.47
    who
    0.42
    Who
    0.39
    åĮħæĭ¬
    0.39
     includ
    0.38
     painters
    0.38
     includes
    0.37
     artists
    0.37
     Who
    0.37
     politicians
    0.37
    Activations Density 0.912%

    No Known Activations