© 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. 22-GEMMASCOPE-RES-131K
    7. 12764
    Prev
    Next
    INDEX
    Explanations

    css units px em

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

    The neuron fires on numeric literal → unit tokens (e.g. “18 px”, “600 px”, “3 em”)—i.e. it detects size/dimension values in code.

    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
    nend
    -1.04
    opie
    -0.97
    ិ
    -0.96
    ørte
    -0.94
    prak
    -0.93
    umente
    -0.93
    burgers
    -0.91
    雳
    -0.91
    gaming
    -0.91
     univerz
    -0.91
    POSITIVE LOGITS
    万円
    0.96
    꾼
    0.94
     indestructible
    0.86
     What
    0.85
    繰り返
    0.85
    high
    0.85
    RUP
    0.84
    尤其是
    0.83
    떄
    0.82
     besonders
    0.81
    Activations Density 0.003%

    No Known Activations