© 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. 22-GEMMASCOPE-RES-131K
    7. 15880
    Prev
    Next
    INDEX
    Explanations

    programming keywords and syntax

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

    The neuron fires most strongly on tokens that are names of API calls or code‐level commands (e.g. function, subprocess, macro or assertion invocation identifiers).

    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
    ということです
    -1.28
    strahlung
    -1.16
    取り組
    -1.14
    方針
    -1.14
     Fassung
    -1.13
    という感じ
    -1.13
    しては
    -1.12
    態度
    -1.11
    儀式
    -1.11
    れません
    -1.10
    POSITIVE LOGITS
    あー
    1.29
    俺は
    1.26
    ことも
    1.26
    清晨
    1.23
    ほら
    1.19
    %",
    1.19
    様な
    1.16
    ṕ
    1.14
    あぁ
    1.14
     sillä
    1.11
    Activations Density 0.007%

    No Known Activations