© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 35354
    Prev
    Next
    INDEX
    Explanations

    St. Patrick’s Day, students learn

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

    The neuron is primarily activated by list‐style or enumerative formatting elements—numerals, punctuation marks (commas, hyphens, periods), and other tokens used to separate or enumerate features.

    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
    %</
    -0.95
    yeron
    -0.93
    .';
    -0.87
     }:
    -0.86
     '\\;'
    -0.84
    *</
    -0.82
     "];
    -0.81
    cześ
    -0.80
    rure
    -0.80
    wij
    -0.80
    POSITIVE LOGITS
    írus
    0.92
     compassion
    0.85
     gigantic
    0.85
     createNew
    0.84
    께
    0.83
     sentimento
    0.82
    Zul
    0.81
    正しい
    0.81
     robusto
    0.81
    たぶん
    0.80
    Activations Density 0.006%

    No Known Activations