© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 3567
    Prev
    Next
    INDEX
    Explanations

    This neuron strongly fires on commas that follow sentence-initial discourse markers (e.g. “But,” “And,” “However,”), i.e. punctuation marking a transitional shift at the start of a clause.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ん
    0.51
    i
    0.47
     は
    0.45
     ให้
    0.42
     بيكون
    0.42
    ー
    0.42
    都是
    0.41
     を
    0.41
     是
    0.41
    z
    0.40
    POSITIVE LOGITS
     analyzers
    0.55
     adolescents
    0.47
     individuals
    0.46
     coworkers
    0.46
     immigrants
    0.44
     subgroups
    0.44
     outliers
    0.44
     perceiving
    0.44
     deliberations
    0.43
     acuity
    0.43
    Activations Density 0.000%

    No Known Activations