© 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. 15358
    Prev
    Next
    INDEX
    Explanations

    analyze and data

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

    The neuron activates on occurrences of the verb “analyze” (and its morphological variants like “analyzed,” “analyzing,” “analyzes”), i.e., words referring to the action of analysis.

    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
    ’
    -3.83
     Of
    -2.63
    ”;
    -2.58
    ”,
    -2.52
     Our
    -2.41
    }\
    -2.36
     mediocre
    -2.34
     quintessential
    -2.31
     That
    -2.23
     obviamente
    -2.19
    POSITIVE LOGITS
    there
    2.75
    just
    2.56
    其妙
    2.50
    (￣
    2.45
    柃
    2.44
     maquillage
    2.41
    読了
    2.33
    结语
    2.31
    on
    2.31
     摄影
    2.28
    Activations Density 0.017%

    No Known Activations