© 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. 26433
    Prev
    Next
    INDEX
    Explanations

    section headers or introductions

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

    This neuron specializes in spotting numeric literals (e.g. measurements, percentages, digits) in code/CSS.

    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
    がこの
    -1.00
     aquest
    -0.99
     sooo
    -0.98
    模特
    -0.97
    ACIO
    -0.95
    fatter
    -0.94
     décider
    -0.94
     nooit
    -0.94
    CARACTERIST
    -0.94
     whether
    -0.93
    POSITIVE LOGITS
     more
    1.15
     only
    1.13
     instead
    1.10
     just
    1.00
     less
    0.93
    Take
    0.92
    Instead
    0.90
     essential
    0.89
     mú
    0.88
     almost
    0.87
    Activations Density 0.003%

    No Known Activations