© 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. 22-GEMMASCOPE-RES-131K
    4. 98424
    Prev
    Next
    INDEX
    Explanations

    often followed by specific nouns

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

    The main thing this neuron does is detect very common function words (articles, prepositions, conjunctions like “the,” “of,” “to,” “or,” “a,” “an,” etc.).

    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
     sicherlich
    -1.29
    每一次
    -1.28
     hierfür
    -1.26
     diesmal
    -1.23
    但也
    -1.22
     erneu
    -1.20
     sämtliche
    -1.18
     insane
    -1.18
    比べると
    -1.17
     promoção
    -1.16
    POSITIVE LOGITS
     an
    1.75
     by
    1.72
     "
    1.64
     in
    1.61
     “
    1.57
     tzw
    1.55
     one
    1.52
     „
    1.51
     on
    1.43
     with
    1.43
    Activations Density 0.064%

    No Known Activations