© 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. 65374
    Prev
    Next
    INDEX
    Explanations

    date and identifiers

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

    The neuron activates on long runs of underscore or other repeated punctuation characters used as visual separators or dividers in the text.

    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
     only
    -1.38
     ONLY
    -1.21
    برای
    -1.15
     although
    -1.11
     to
    -1.11
     Using
    -1.09
     JUST
    -1.08
     even
    -1.06
     computa
    -1.06
    Using
    -1.05
    POSITIVE LOGITS
     calefacción
    1.09
    的在
    1.00
     rive
    0.96
    垮
    0.96
    にとっては
    0.95
    都好
    0.95
     materiału
    0.94
     aDecoder
    0.93
     européens
    0.92
    ,'
    0.92
    Activations Density 0.010%

    No Known Activations