© 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. 1941
    Prev
    Next
    INDEX
    Explanations

    This neuron fires strongly on mathematical expressions—numbers, variables, and equation notation.

    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
    研修
    0.78
     Ambassadors
    0.78
     Geschä
    0.78
     Bazaar
    0.77
     Einige
    0.77
     Política
    0.76
     Politik
    0.76
     彼
    0.76
     Muchos
    0.75
     Gente
    0.75
    POSITIVE LOGITS
     equation
    1.46
     уравнение
    1.34
     equations
    1.30
     substituting
    1.26
    Substituting
    1.22
     eqn
    1.22
     уравнения
    1.21
     Substituting
    1.20
    equation
    1.16
     Equation
    1.16
    Activations Density 0.050%

    No Known Activations