© 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. 22-GEMMASCOPE-RES-131K
    7. 72155
    Prev
    Next
    INDEX
    Explanations

    not followed by verb

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

    The neuron strongly activates on negation tokens (e.g. “not,” “can’t,” “don’t,” and the ’t suffix), i.e. markers of negative polarity.

    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
     éléments
    -1.07
     الاطلاع
    -1.04
     ELEMENTS
    -1.02
     zcela
    -1.02
     complètement
    -1.00
    negara
    -0.98
    mettre
    -0.98
    }}}{
    -0.98
     verminder
    -0.96
     défis
    -0.95
    POSITIVE LOGITS
    Nieuwe
    1.21
     even
    1.10
     ostensibly
    1.09
    лтемелер
    1.05
     impressive
    1.05
    ække
    1.04
     །
    1.04
    Hör
    1.02
    Nieuw
    1.01
     for
    0.99
    Activations Density 0.009%

    No Known Activations