© 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. 26017
    Prev
    Next
    INDEX
    Explanations

    table formatting

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

    The neuron strongly activates on tokens consisting entirely of repeated characters (especially “=” lines) or pure digit strings—that is, visual separators and numeric identifiers.

    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
     when
    -1.25
     only
    -1.20
    きは
    -1.06
    when
    -1.05
     where
    -1.00
    pirazione
    -0.91
     specifically
    -0.91
     through
    -0.90
    chaise
    -0.90
     of
    -0.87
    POSITIVE LOGITS
     другое
    1.16
    呶
    1.12
     другую
    1.06
     paneles
    1.05
    Ingrédients
    1.04
     suiv
    1.02
    쐈
    1.02
    ftar
    1.00
    ÈNE
    1.00
     selve
    0.99
    Activations Density 0.002%

    No Known Activations