© 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. 10-GEMMASCOPE-RES-131K
    4. 41868
    Prev
    Next
    INDEX
    Explanations

    words that are said or read

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

    The neuron flags tokens that signal quoted or reported text—especially verbs like “said,” “read,” “says,” “reading” and accompanying quotation marks.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     it
    -2.17
     when
    -2.06
     with
    -1.99
     these
    -1.78
     should
    -1.55
     []:
    -1.55
    These
    -1.55
     because
    -1.54
     what
    -1.53
     there
    -1.52
    POSITIVE LOGITS
     a
    2.05
    涠
    1.67
     “
    1.59
    ⎆
    1.55
    “
    1.54
    
    1.52
     primar
    1.52
     resul
    1.48
    methoden
    1.48
     signif
    1.46
    Activations Density 0.052%

    No Known Activations