© 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. 22-GEMMASCOPE-RES-131K
    7. 54188
    Prev
    Next
    INDEX
    Explanations

    punctuation then auxiliary verbs

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

    The neuron is detecting long dashes (“--”) or other extended hyphen‐style punctuation used to set off parenthetical or aside material.

    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
     melap
    -1.73
    teilte
    -1.66
     بودند
    -1.66
    danke
    -1.63
     mengucap
    -1.62
    banez
    -1.60
     bluz
    -1.57
     菇
    -1.55
     récol
    -1.52
    ~”
    -1.50
    POSITIVE LOGITS
     is
    3.33
     are
    3.28
     was
    3.13
     will
    2.55
     would
    2.45
     may
    2.41
     have
    2.39
     can
    2.31
     should
    2.28
     could
    2.25
    Activations Density 0.103%

    No Known Activations