© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 32848
    Prev
    Next
    INDEX
    Explanations

    first letters followed by suffixes

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

    The neuron strongly activates on isolated single-letter or very short uppercase tokens (e.g. “A,” “C,” “L,” “F,” “B,” etc.) typically used as initials or abbreviations.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     Sorting
    -1.19
     tragé
    -1.07
    mô
    -1.06
     asisten
    -1.05
     plafon
    -1.05
     dépens
    -1.05
     microfibra
    -1.04
     useCallback
    -1.02
     seraient
    -1.00
     maquilla
    -1.00
    POSITIVE LOGITS
     but
    1.25
     will
    1.22
     as
    1.17
     because
    1.16
    没关系
    1.05
     those
    1.00
    最后一
    0.97
    O
    0.94
     may
    0.93
    没问题
    0.92
    Activations Density 0.006%

    No Known Activations