© 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. 10-GEMMASCOPE-RES-131K
    7. 80677
    Prev
    Next
    INDEX
    Explanations

    have/had + past participle verbs

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

    The neuron selectively activates on core English auxiliary and past-tense/past-participle verbs (e.g. have, had, been, joined, created, patched, protected).

    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
    ܖ
    -2.09
     pè
    -1.97
    Hanya
    -1.95
     Père
    -1.92
     auffi
    -1.88
     Ministre
    -1.80
     sentier
    -1.80
     estadounid
    -1.79
    Siapa
    -1.75
    Setiap
    -1.75
    POSITIVE LOGITS
     that
    2.20
     or
    2.14
     which
    2.03
     from
    2.02
     to
    1.92
     [
    1.77
     Or
    1.75
     we
    1.72
     In
    1.67
     You
    1.61
    Activations Density 0.024%

    No Known Activations