© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 108846
    Prev
    Next
    INDEX
    Explanations

    swimming, water, pool

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

    This neuron primarily detects numeric digit tokens, i.e. parts of numbers (years, event numbers, measurements) in the text.

    oai_token-act-pair · o4-miniTriggered 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
    🎣
    -0.82
    мелер
    -0.78
     motocross
    -0.78
     chitar
    -0.72
    herbe
    -0.70
    Succeed
    -0.70
     Sua
    -0.69
    🎾
    -0.69
    Stderr
    -0.68
    README
    -0.67
    POSITIVE LOGITS
     water
    1.42
    🤽
    1.38
    water
    1.30
     Water
    1.19
     WATER
    1.14
    Water
    1.10
     attacker
    1.05
     pool
    1.02
    WATER
    1.02
     chlorine
    0.98
    Activations Density 0.010%

    No Known Activations