© 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. 22-GEMMASCOPE-RES-131K
    7. 14366
    Prev
    Next
    INDEX
    Explanations

    Gabriel Newton

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

    The neuron activates on numeric literals—any tokens composed of digits (e.g., numbers in code, dates, measurements).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
    nutie
    -0.88
     포함
    -0.86
     instalaciones
    -0.85
    خخ
    -0.85
    Kde
    -0.83
    军的
    -0.83
     solicitar
    -0.82
     fortale
    -0.82
    予定
    -0.82
     đề
    -0.82
    POSITIVE LOGITS
    이터
    0.98
    SQLiteDatabase
    0.91
    stylesheet
    0.88
    zenten
    0.85
    МІ
    0.85
    躍
    0.84
     மற்ற
    0.83
    셀
    0.83
    čená
    0.83
    DATABASES
    0.82
    Activations Density 0.004%

    No Known Activations