© 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. 16145
    Prev
    Next
    INDEX
    Explanations

    special characters and code delimiters

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

    The neuron is keyed to backslash‐style tokens (e.g. “\” sequences), effectively spotting LaTeX‐style commands.

    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
    の人に
    -1.05
    횃
    -0.95
    ticularly
    -0.88
    ヌーピー
    -0.86
    ˮ
    -0.84
     metálica
    -0.84
     руку
    -0.83
     ایرانی
    -0.82
     travaillent
    -0.82
    plante
    -0.81
    POSITIVE LOGITS
     denial
    0.79
    }_{\
    0.75
    佻
    0.74
     widespread
    0.73
    願い
    0.73
     złoż
    0.72
    bParam
    0.72
    と思っていた
    0.71
    램
    0.71
    iben
    0.71
    Activations Density 0.006%

    No Known Activations