© 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. 99341
    Prev
    Next
    INDEX
    Explanations

    quarterly

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

    The neuron specifically activates on the word “quarterly,” flagging mentions of that periodicity (e.g. journal or event frequency).

    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
     The
    -3.69
     It
    -3.06
     Ideally
    -2.92
     berühmten
    -2.88
     Для
    -2.83
     --
    -2.77
     However
    -2.72
     Again
    -2.70
     Which
    -2.67
     Should
    -2.67
    POSITIVE LOGITS
    焢
    3.16
    )
    2.83
    just
    2.67
     menangis
    2.48
     土産
    2.45
    掮
    2.34
    ⑊
    2.34
    Even
    2.31
    there
    2.30
     тебя
    2.30
    Activations Density 0.005%

    No Known Activations