© 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. 10-GEMMASCOPE-RES-131K
    7. 58570
    Prev
    Next
    INDEX
    Explanations

    mathematical notation and names

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

    The neuron fires strongly on LaTeX “\mathbf” tokens, i.e. it detects occurrences of bold‐face math markup.

    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
    ’
    -3.80
     ONLY
    -2.64
    of
    -2.48
     exuberant
    -2.41
    üşonlu
    -2.38
     JUST
    -2.31
    s
    -2.28
     blatantly
    -2.28
     ingrained
    -2.27
    :
    -2.23
    POSITIVE LOGITS
    {
    3.28
    ౧
    3.02
    盉
    2.83
    泐
    2.78
    龗
    2.50
    顚
    2.48
     Infór
    2.45
    ჻
    2.45
     jezt
    2.44
    荅
    2.44
    Activations Density 0.002%

    No Known Activations