© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 95425
    Prev
    Next
    INDEX
    Explanations

    titles ending in twelve

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

    This neuron detects numeric tokens (multi-digit numbers, years, and statistical counts).

    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
     removes
    -1.07
     immediately
    -1.04
     encourages
    -1.03
     encourage
    -1.01
     crí
    -0.99
     an
    -0.96
     takes
    -0.96
     thinks
    -0.95
     reduces
    -0.95
     it
    -0.94
    POSITIVE LOGITS
     tournant
    1.06
    STOT
    1.05
    Poznám
    1.04
     capitaine
    1.00
     lesquelles
    0.99
     Kindern
    0.98
     STARTING
    0.97
     الشاشة
    0.97
     poème
    0.97
     quelquefois
    0.96
    Activations Density 0.003%

    No Known Activations