© 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. 41853
    Prev
    Next
    INDEX
    Explanations

    aws, extend and bind

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

    The neuron fires on programming‐language tokens (like template‐literal backticks, import paths, code identifiers, method names, etc.), i.e. it detects snippets of code rather than ordinary prose.

    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
    );
    -1.84
     categorized
    -1.77
     differentiates
    -1.73
     moderately
    -1.69
     subtly
    -1.66
     interestingly
    -1.66
     makes
    -1.64
     unequivocally
    -1.61
    形成了
    -1.60
     suggests
    -1.55
    POSITIVE LOGITS
    ↵
    1.98
    ⩖
    1.71
     the
    1.70
    yeah
    1.68
    鲩
    1.64
    </u>
    1.57
    訫
    1.54
     Dinas
    1.52
     ストライプ
    1.51
     tento
    1.45
    Activations Density 0.002%

    No Known Activations