© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 73255
    Prev
    Next
    INDEX
    Explanations

    words or phrases following punctuation

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

    The neuron is detecting sentence boundaries—spiking on end-of-sentence punctuation (and quotes) and on the first word of the next 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_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
     ДЕ
    -1.02
     aihe
    -0.96
     tohto
    -0.90
     belge
    -0.89
    gencias
    -0.89
    bege
    -0.89
    的同时
    -0.88
    줄
    -0.88
     acestui
    -0.88
     AVENUE
    -0.87
    POSITIVE LOGITS
     then
    1.23
     soon
    1.21
     completed
    1.05
     finished
    1.03
     after
    0.96
     proceed
    0.95
     anyway
    0.94
     finally
    0.93
     later
    0.93
     when
    0.93
    Activations Density 0.012%

    No Known Activations