© 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. 64853
    Prev
    Next
    INDEX
    Explanations

    ON ; invention are

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

    The neuron fires most strongly on code punctuation that delimits statements (e.g. semicolons, braces, and statement‐terminating tokens).

    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
     matériaux
    -1.02
     бампер
    -1.01
     automó
    -1.00
     différentes
    -1.00
     condomínio
    -0.97
    rii
    -0.96
     UserDao
    -0.96
    pastebin
    -0.96
    oubtedly
    -0.94
     différents
    -0.94
    POSITIVE LOGITS
     started
    1.34
     and
    1.30
     began
    1.18
     start
    1.12
     had
    1.10
     on
    1.10
     for
    1.10
     time
    1.09
     starts
    1.08
     begins
    1.08
    Activations Density 0.002%

    No Known Activations