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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 70992
    Prev
    Next
    INDEX
    Explanations

    other languages, other values

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

    The neuron fires on numeric literals (especially multi‐digit and decimal numbers) embedded in code or formulas.

    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
    papel
    -0.98
    ˵
    -0.90
     what
    -0.85
    ごちそうさまでした
    -0.85
     núi
    -0.84
     young
    -0.84
    おわりに
    -0.83
    travail
    -0.83
    teatro
    -0.78
     that
    -0.78
    POSITIVE LOGITS
    nisse
    0.85
     أخرى
    0.82
    他の
    0.81
     other
    0.80
    yet
    0.79
    non
    0.79
     andere
    0.78
    別の
    0.78
    hof
    0.78
    чивать
    0.77
    Activations Density 0.023%

    No Known Activations