© 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. 22-GEMMASCOPE-RES-131K
    4. 39628
    Prev
    Next
    INDEX
    Explanations

    code blocks and keywords

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

    The neuron fires on standalone code‐oriented lines—especially comment lines, section headings or other non-natural-language markers (e.g. “// …”, “<?php …?>”, “-- FILE --”, semicolon-terminated declarations)—i.e. structural/formatting lines in code examples.

    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
    我が家の
    -1.24
     preva
    -1.18
     表紙
    -1.17
    ほら
    -1.13
    我が家
    -1.10
     огур
    -1.10
     vín
    -1.09
     gratuite
    -1.09
    裃
    -1.08
     hias
    -1.07
    POSITIVE LOGITS
     and
    1.13
     F
    0.96
     AND
    0.91
     B
    0.87
     ......
    0.86
    jsonArray
    0.86
     tells
    0.84
     has
    0.81
     E
    0.80
     みた
    0.80
    Activations Density 0.052%

    No Known Activations