© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 48266
    Prev
    Next
    INDEX
    Explanations

    French titles and phrases

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

    The neuron fires on single-word heading or title tokens (e.g. standalone labels like “bon,” “Chant,” “belle,” “Grandes,” “常识”) that serve as section‐ or work‐titles rather than running text.

    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
    圌
    -2.61
    鲩
    -2.48
    驺
    -2.45
    臜
    -2.42
    咵
    -2.39
    ౧
    -2.33
    ⎘
    -2.31
    踯
    -2.28
    苤
    -2.28
    鼍
    -2.28
    POSITIVE LOGITS
     You
    2.86
    You
    2.72
    This
    2.69
    t
    2.64
    ."
    2.38
    );
    2.33
    It
    2.23
     For
    2.22
     sebagainya
    2.20
    。
    2.19
    Activations Density 0.003%

    No Known Activations