© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 16467
    Prev
    Next
    INDEX
    Explanations

    as/is/am/are/were/has/did

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

    The neuron strongly activates on personal pronouns (I, he, they, etc.) and copular verbs (am, is, were, etc.), effectively marking subject–to-be constructions.

    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
    tismo
    -1.26
    お父
    -1.17
    这一幕
    -1.16
    我が家の
    -1.16
    籏
    -1.15
    峮
    -1.15
    そういえば
    -1.12
    スポンサーリンク
    -1.10
    quelize
    -1.09
    はコチラ
    -1.09
    POSITIVE LOGITS
     to
    1.37
     of
    1.24
    了承
    1.22
     than
    1.21
     being
    1.15
    有任何
    1.13
     it
    1.13
     any
    1.07
     eines
    1.05
     they
    1.02
    Activations Density 0.016%

    No Known Activations