© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 95675
    Prev
    Next
    INDEX
    Explanations

    integer numeric types

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

    The neuron strongly responds to numeric tokens—either literal numbers or identifiers related to numeric types (e.g. “int,” “integer,” “decimal,” and actual digit sequences).

    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
     các
    -1.09
    C
    -1.07
     those
    -1.07
     all
    -1.06
     …
    -1.05
    那些
    -1.05
     including
    -1.03
     [
    -1.02
     such
    -0.97
     รวม
    -0.97
    POSITIVE LOGITS
     rémun
    1.42
    uteurs
    1.41
     berücksich
    1.29
     trône
    1.22
     pokoj
    1.22
     jednod
    1.22
     čís
    1.19
     geçti
    1.18
     ministerio
    1.17
     sabid
    1.16
    Activations Density 0.018%

    No Known Activations