© 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. 5005
    Prev
    Next
    INDEX
    Explanations

    css flexbox properties

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

    This neuron consistently fires on the first “content” word of a new clause or sentence (e.g. the initial non‐stopword token that begins a numbered item, sentence, or section).

    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
     ombil
    0.93
     цієї
    0.88
     aquesta
    0.87
     aucun
    0.84
    トル
    0.84
     imaju
    0.84
    這
    0.84
     enregist
    0.83
     défin
    0.83
    hatikan
    0.83
    POSITIVE LOGITS
    ле
    0.69
    </h2>
    0.68
    is
    0.66
    </blockquote>
    0.66
    he
    0.66
    hon
    0.65
    Union
    0.64
     Corporation
    0.63
     mets
    0.63
    ová
    0.63
    Activations Density 0.002%

    No Known Activations