© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 64682
    Prev
    Next
    INDEX
    Explanations

    Japanese patent applications

    np_acts-logits-general · gemini-2.5-flash-lite

    The neuron chiefly fires on the token “Japanese,” i.e. instances of the word Japanese in patent‐style text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     rijden
    -0.85
     obec
    -0.83
     ketika
    -0.81
    льной
    -0.80
     Mexico
    -0.79
     étaient
    -0.76
     salieron
    -0.75
    áctica
    -0.75
     akku
    -0.74
     heti
    -0.74
    POSITIVE LOGITS
     Becomes
    0.92
    レア
    0.86
    izu
    0.84
     remains
    0.84
    vám
    0.80
     Remains
    0.80
     Deleg
    0.77
     بالفعل
    0.76
    不好
    0.76
    Series
    0.75
    Activations Density 0.001%

    No Known Activations