© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 75326
    Prev
    Next
    INDEX
    Explanations

    google internal import

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

    This neuron fires on code tokens that are names of software projects, libraries, or packages (e.g. “minecraft,” “google,” “apache,” “lodash,” “eclipse”).

    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
     páns
    -1.05
     никакого
    -1.04
    beige
    -1.02
     الرغم
    -1.02
    つづく
    -1.00
     provincie
    -1.00
     smrti
    -1.00
    albeit
    -1.00
     compter
    -0.99
    shadowOffset
    -0.98
    POSITIVE LOGITS
     then
    1.19
    Then
    1.01
    said
    1.00
    </strong>
    0.98
     says
    0.98
     is
    0.98
    1
    0.97
     say
    0.96
     need
    0.94
    set
    0.93
    Activations Density 0.001%

    No Known Activations