© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 15146
    Prev
    Next
    INDEX
    Explanations

    paths, calculations, higher

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

    The neuron strongly activates on relatively long, uncommon content words (e.g. abstract or technical nouns of seven or more letters).

    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
     мы
    0.71
    мы
    0.65
    デジタル
    0.64
    Character
    0.63
     {
    0.61
    itosan
    0.61
    они
    0.60
     우리는
    0.59
    旱
    0.59
    character
    0.59
    POSITIVE LOGITS
     menjalankan
    0.74
     unresponsive
    0.74
     oran
    0.71
     .)
    0.70
     demanding
    0.68
    बम
    0.67
     ENGLAND
    0.66
     dirs
    0.65
     INTERNATIONAL
    0.64
     பார்
    0.63
    Activations Density 0.000%

    No Known Activations