© 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-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 6791
    Prev
    Next
    INDEX
    Explanations

    classifying data

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

    The neuron fires strongly on interface or metadata tokens—things like “Search”, “for”, HTML attribute names (e.g. rel=), page labels in other scripts (“صفحه”), and similar UI/navigation cues.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Española
    0.93
    ຟ
    0.83
    וס
    0.82
    lana
    0.82
     ideas
    0.80
    יות
    0.79
    łod
    0.79
     Óscar
    0.78
     Realtors
    0.77
     kilobytes
    0.77
    POSITIVE LOGITS
     государственный
    0.71
    怎樣
    0.70
    лло
    0.70
     나오고
    0.69
    у
    0.69
    announce
    0.68
    ocus
    0.66
    warn
    0.66
    examine
    0.66
     два
    0.66
    Activations Density 0.001%

    No Known Activations