© 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. 17
    Prev
    Next
    INDEX
    Explanations

    while subject verb

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

    The neuron is essentially a comma detector – it fires strongly on comma tokens (and similar punctuation that marks clause breaks).

    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
     kompati
    -1.46
     phénom
    -1.41
    몄
    -1.38
     signifikan
    -1.36
     konkre
    -1.36
     rozs
    -1.35
     mystérie
    -1.34
     múl
    -1.33
    ⚢
    -1.33
     étroite
    -1.31
    POSITIVE LOGITS
     not
    1.98
     at
    1.52
     or
    1.52
     from
    1.45
     what
    1.42
     on
    1.36
     just
    1.25
     of
    1.24
     nor
    1.23
     except
    1.23
    Activations Density 0.041%

    No Known Activations