© 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. 10-GEMMASCOPE-RES-131K
    7. 44627
    Prev
    Next
    INDEX
    Explanations

    once followed by past states

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

    The main thing this neuron does is detect the adverb “once” indicating a past occurrence.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    2
    -1.73
    0
    -1.58
    9
    -1.55
     liknande
    -1.52
    bundet
    -1.51
     forbin
    -1.49
     område
    -1.48
    7
    -1.48
     {},
    -1.47
     本店
    -1.45
    POSITIVE LOGITS
    ’
    1.69
     églises
    1.58
     моего
    1.54
     erroneously
    1.50
     mistakenly
    1.50
     disdain
    1.49
    If
    1.48
     be
    1.48
     мужа
    1.48
    軻
    1.45
    Activations Density 0.013%

    No Known Activations