© 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-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 8718
    Prev
    Next
    INDEX
    Explanations

    Donetsk People's Republic

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

    The neuron flags lines that look like image‐ or media‐item titles/headings (e.g. “Illustration …,” “Drawing …,” “Photo …,” “The conference …,” “Picture …”)—the kind of caption or credit heading you’d see before a “from X, source: Y” attribution.

    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
    UCE
    0.85
    களில்
    0.83
    нача
    0.83
    酒吧
    0.83
    지난
    0.82
    刎
    0.81
    가
    0.80
    我和
    0.78
    我在
    0.77
     модер
    0.77
    POSITIVE LOGITS
    ÇÕES
    0.79
     Portanto
    0.77
     Cũng
    0.75
     Meinung
    0.71
    &
    0.71
     Blur
    0.70
    тивных
    0.70
    вшийся
    0.70
     THAT
    0.70
     Vậy
    0.70
    Activations Density 0.000%

    No Known Activations