© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 66278
    Prev
    Next
    INDEX
    Explanations

    negation with do

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

    The neuron detects occurrences of the auxiliary verb “do” (and its past‐tense form “did” and third‐person singular “does”).

    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
     rése
    -1.91
     honnête
    -1.80
     écrivain
    -1.69
     journaliste
    -1.64
    纐
    -1.59
    縵
    -1.58
    杙
    -1.56
     piden
    -1.55
     exercer
    -1.54
     exé
    -1.53
    POSITIVE LOGITS
    ’
    1.70
    </h2>
    1.54
    蔔
    1.52
     however
    1.44
    하지
    1.44
    </h4>
    1.42
     not
    1.40
     could
    1.39
    </strong>
    1.37
     presenti
    1.37
    Activations Density 0.024%

    No Known Activations