© 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-3-27B-IT
    3. 10-GEMMASCOPE-2-TRANSCODER-262K
    4. 45841
    Prev
    Next
    INDEX
    Explanations

    /**This neuron is associated with JSDoc comments and their structure. The `MAX_ACTIVATING_TOKENS` list overwhelmingly features 'doc' and 'oc', which are key components of a JSDoc declaration (often starting like `/**`). The `TOKENS_AFTER_MAX_ACTIVATING_TOKEN` list further supports this by showing tokens like 'for', 'Comments', and 'comment', which are common within JSDoc blocks to describe parameters, the function's purpose, or general notes. The `TOP_ACTIVATING_TEXTS` are filled with examples of code that use JSDoc comments for documentation (`jsdoc for function`, `JSDoc comments`, `Docstring`, `/** @type {import('next').NextConfig} */`, `JSDoc-style comment`).Therefore, the neuron seems to activate when it encounters or predicts JSDoc documentation.Explanation: JSDoc comments and documentation

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/transcoder_all/layer_10_width_262k_l0_small_affine
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Przy
    0.57
     Beziehungen
    0.57
     леген
    0.54
    RatingDiff
    0.53
    ♲
    0.53
    স্পতিবার
    0.52
     Enth
    0.52
     hurts
    0.52
     Beziehung
    0.52
     éxitos
    0.52
    POSITIVE LOGITS
    ا
    0.71
    作り
    0.62
    ‌های
    0.58
    -
    0.57
    ższ
    0.56
    ಗೊಳ
    0.56
    राज
    0.54
     policial
    0.54
    ्तीय
    0.53
    ᱔
    0.53
    Activations Density 0.000%

    No Known Activations