© 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. 10-GEMMASCOPE-RES-131K
    7. 22866
    Prev
    Next
    INDEX
    Explanations

    specific determiners and common nouns

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

    This neuron responds to structural headings or other text presented in all-caps (e.g. section titles, figure labels, “LOCKER ROOM,” “GENERAL,” “POWER,” etc.).

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     colonnes
    -1.30
     envío
    -1.12
     burada
    -1.12
    ']")
    -1.12
    ihnach
    -1.12
     arbres
    -1.10
     recevez
    -1.09
     secado
    -1.08
    を作って
    -1.08
     visuales
    -1.06
    POSITIVE LOGITS
     who
    1.70
     in
    1.67
     people
    1.63
     if
    1.55
     those
    1.52
     to
    1.43
    Not
    1.42
     you
    1.41
    useNewUrlParser
    1.36
     defStyle
    1.30
    Activations Density 0.271%

    No Known Activations