© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 2972
    Prev
    Next
    INDEX
    Explanations

    explaining a formula

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

    The neuron activates on numeric tokens—especially multi‐digit and decimal numbers—i.e. it picks out numbers in the text.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Ninety
    0.58
     Biele
    0.54
     Zahlen
    0.52
     Instantiate
    0.52
    翊
    0.52
    vní
    0.51
    NUMBER
    0.50
    契
    0.50
     Number
    0.50
     NUMBER
    0.50
    POSITIVE LOGITS
    formula
    0.96
     formula
    0.93
     simple
    0.89
     formulas
    0.89
     semplice
    0.88
    formulas
    0.88
     Formula
    0.87
    Formula
    0.86
     fórmula
    0.84
     semplici
    0.84
    Activations Density 0.002%

    No Known Activations