© 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. 102065
    Prev
    Next
    INDEX
    Explanations

    discourse starter "so"

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

    The neuron fires strongly on sentence-initial “So,” i.e. the discourse marker “So” at the start of a clause or sentence.

    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
     and
    -2.44
     has
    -2.31
     will
    -2.05
     is
    -2.02
     mientras
    -1.88
    に
    -1.82
     incluyendo
    -1.75
    it
    -1.72
    哲学
    -1.68
    ”),
    -1.65
    POSITIVE LOGITS
     habile
    1.90
    螃
    1.86
    provide
    1.66
    それとも
    1.63
     \%}$
    1.62
     nonatomic
    1.62
    othed
    1.62
    ܆
    1.59
     desenli
    1.59
     hebben
    1.59
    Activations Density 0.031%

    No Known Activations