© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 67310
    Prev
    Next
    INDEX
    Explanations

    HTML structure, components, classes

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

    The neuron fires on the structural tokens of embedded JSX/HTML/XML-like markup—essentially detecting the “<” tag openings (and associated JSX tag syntax) in the code.

    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
     ファイル
    -1.63
     neón
    -1.48
     Questions
    -1.37
     niebla
    -1.35
     sagt
    -1.35
    thello
    -1.30
     dieß
    -1.30
     zrobić
    -1.29
     cuarzo
    -1.29
    OREM
    -1.28
    POSITIVE LOGITS
     with
    1.80
     despite
    1.68
    残り
    1.38
    我对
    1.34
    それとも
    1.30
     bästa
    1.28
     anum
    1.26
    よくある
    1.24
    实习
    1.23
    ”,
    1.22
    Activations Density 0.005%

    No Known Activations