© 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. 73597
    Prev
    Next
    INDEX
    Explanations

    punctuation followed by evaluation

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

    The neuron is detecting punctuation that marks quoted speech—especially quotation marks, commas, and sentence-ending periods around direct quotes.

    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
     just
    -2.47
     with
    -2.44
     undeniably
    -2.34
     epitom
    -2.25
     both
    -2.25
     two
    -2.16
     ostensibly
    -2.16
     spurred
    -2.13
     only
    -2.11
     several
    -2.11
    POSITIVE LOGITS
     horrid
    2.20
     delightfully
    2.19
     lauded
    2.17
     strikingly
    2.16
     wondrous
    2.16
     ludicrous
    2.14
     astounding
    2.11
     boister
    2.08
     ridiculously
    2.08
     secluded
    2.06
    Activations Density 0.052%

    No Known Activations