© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 105533
    Prev
    Next
    INDEX
    Explanations

    waited leisurely

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

    The neuron fires on the little connective tokens and state-describing words that hold a narrative together—prepositions, conjunctions, transition adverbs, and verbs or nouns that name conditions (e.g. “in,” “of,” “again,” “waited,” “oppression,” “state”).

    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
     to
    -2.05
    њу
    -1.39
     now
    -1.34
    着一个
    -1.30
     начинают
    -1.30
     可愛い
    -1.29
    gilet
    -1.28
     then
    -1.28
    を見せて
    -1.27
    *}{
    -1.25
    POSITIVE LOGITS
     пр
    1.41
    ulario
    1.40
    話
    1.38
     Sudah
    1.37
     mécan
    1.36
     aveva
    1.34
     vitesses
    1.30
     saati
    1.29
                       
    1.28
    istä
    1.27
    Activations Density 0.005%

    No Known Activations