© 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. 104674
    Prev
    Next
    INDEX
    Explanations

    Created, This, Mi, If, Pres

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

    The neuron strongly activates on major punctuation marks (periods or commas) that occur after long, information‐dense spans of text, effectively marking sentence boundaries in extended document passages.

    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
     first
    -1.44
     ilustracji
    -1.39
     then
    -1.37
     only
    -1.33
    puma
    -1.30
     just
    -1.27
     from
    -1.26
     several
    -1.26
    淩
    -1.25
     five
    -1.22
    POSITIVE LOGITS
     offiziell
    1.42
     gravi
    1.33
     unn
    1.29
     Dinas
    1.29
    orm
    1.28
     dee
    1.27
     horrid
    1.27
    vées
    1.26
     myriad
    1.25
     revered
    1.25
    Activations Density 0.050%

    No Known Activations