© 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
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 97820
    Prev
    Next
    INDEX
    Explanations

    ISBNs and publication details

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

    This neuron is highly sensitive to the phrase “Oceanic and Atmospheric,” as in “National Oceanic and Atmospheric Administration.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    Neuron 4 activates for numeric digits, particularly in sequences like years, version numbers, measurements, or identifiers.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    formal or technical labels and names in text, such as titles, organizational/institutional terms, and configuration/parameter identifiers.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    extension
    -0.77
    Extension
    -0.77
     extension
    -0.76
    こっち
    -0.75
     Extension
    -0.73
     nacht
    -0.73
     PRE
    -0.72
     canales
    -0.71
     oko
    -0.70
    Integrated
    -0.70
    POSITIVE LOGITS
     Testa
    0.78
     helix
    0.67
     seize
    0.65
    MIRA
    0.64
     angina
    0.62
     ходить
    0.62
    ！』
    0.62
    captcha
    0.62
    CDR
    0.61
    Gm
    0.60
    Activations Density 0.058%

    No Known Activations