© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 130996
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on long runs of punctuation characters (e.g. sequences of hyphens or asterisks) used as table/list separators or dividers.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    differences between

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     to
    -1.47
     from
    -1.34
     by
    -1.30
     for
    -1.27
    ctober
    -1.15
    СЫ
    -1.12
     you
    -1.11
     I
    -1.09
     September
    -1.08
     Приступљено
    -1.06
    POSITIVE LOGITS
    Pilih
    1.17
    がありますが
    1.14
    Profes
    1.11
     entren
    1.11
    Missi
    1.10
     erweiter
    1.08
    nappropriate
    1.06
    expensive
    1.05
     reutiliz
    1.04
     plateado
    1.03
    Activations Density 0.002%

    No Known Activations