© 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. 34-GEMMASCOPE-RES-131K
    7. 25988
    Prev
    Next
    INDEX
    Explanations

    HTTP, dissolved oxygen, built by, contraction of

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

    The neuron fires on tokens that are (or contain) numeric strings—e.g. version numbers, digit sequences, or other number‐heavy identifiers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    書店
    -0.76
    addAction
    -0.73
    仟
    -0.72
     QC
    -0.69
     lunatic
    -0.67
     considérable
    -0.67
    гка
    -0.67
    méno
    -0.67
    зон
    -0.66
     floodplain
    -0.66
    POSITIVE LOGITS
    haz
    0.70
    針對
    0.67
    zzles
    0.66
     PSA
    0.65
     Infan
    0.64
    顯
    0.64
    fitrión
    0.63
     accessToken
    0.63
     Pó
    0.62
    рования
    0.61
    Activations Density 0.076%

    No Known Activations