© 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. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 129051
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on numeric tokens and number-like patterns (digits, subscripts, decimal numbers) in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    last element

    np_acts-logits-general · gemini-2.5-flash-lite
    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
    Два
    -0.88
    Продол
    -0.84
    annon
    -0.82
     svart
    -0.82
     inteligen
    -0.82
     corrom
    -0.81
    ܓ
    -0.81
    Sinopsis
    -0.79
     menghasilkan
    -0.79
     Fortaleza
    -0.79
    POSITIVE LOGITS
     max
    1.19
     Max
    1.03
     last
    1.02
     конце
    0.98
     MAX
    0.94
     maximum
    0.93
     min
    0.88
     available
    0.87
     послед
    0.87
     kẻ
    0.87
    Activations Density 0.045%

    No Known Activations