© 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. 31-TRANSCODER-HP
    7. 18399
    Prev
    Next
    INDEX
    Explanations

    Poland

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    mwhanna/qwen3-4b-transcoders/layer_31.safetensors
    Prompts (Dashboard)
    16,384 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    163,840
    Data Type
    float32
    Hook Name
    blocks.31.mlp.hook_in
    Architecture
    transcoder
    Context Size
    8,192
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Conrad
    -0.29
    rogram
    -0.26
    éªħ
    -0.25
    หล
    -0.25
    å½ĵäºĭ人
    -0.24
    è¯ģ
    -0.24
    .widget
    -0.24
     programmed
    -0.24
    ponent
    -0.24
    æĪIJé¾Ļ
    -0.23
    POSITIVE LOGITS
    ä¸ĬäºĨ
    0.26
    _unset
    0.25
     histograms
    0.25
    illos
    0.24
    OUND
    0.24
     ammon
    0.24
     unlink
    0.24
    åħ¬ç«ł
    0.24
    epend
    0.23
    licable
    0.23
    Activations Density 0.331%

    No Known Activations