© 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. 25952
    Prev
    Next
    INDEX
    Explanations

    mathematical expressions and inequalities

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

    The neuron fires on numeric tokens (digit sequences) in mathematical expressions—i.e. it detects numbers (constants, coefficients, limits, etc.).

    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
    urgie
    -1.01
    REQUENCY
    -0.96
    zens
    -0.94
    newblock
    -0.93
     preguntó
    -0.92
    ContentAlignment
    -0.92
     TabLayout
    -0.90
     Heidegger
    -0.89
     proget
    -0.89
     ViewPager
    -0.89
    POSITIVE LOGITS
     and
    1.01
    ὴν
    1.00
    thre
    0.97
     until
    0.94
     limitado
    0.93
     it
    0.92
     between
    0.91
     finalmente
    0.87
    нную
    0.84
    uertos
    0.83
    Activations Density 0.024%

    No Known Activations