© 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-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 100381
    Prev
    Next
    INDEX
    Explanations

    JSON and HTML tags

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

    the "title" attribute or field in HTML/XML/JSON markup, particularly when followed by an equals sign and quote.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron fires strongly on “title” fields or attributes (e.g. JSON `"title":` or HTML `title="…"`), essentially spotting metadata keys named title and their following separators.

    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
     emphasizes
    -1.81
     helped
    -1.78
     took
    -1.69
    寕
    -1.64
     considerable
    -1.64
     represents
    -1.63
     had
    -1.62
     tries
    -1.59
    家は
    -1.58
     sparked
    -1.56
    POSITIVE LOGITS
    see
    2.14
    are
    2.09
    趿
    1.98
    τα
    1.96
    镡
    1.95
    𝅥
    1.91
     ۱۳
    1.89
     andere
    1.77
    镲
    1.74
    ны
    1.73
    Activations Density 0.017%

    No Known Activations