© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 2583
    Prev
    Next
    INDEX
    Explanations

    hidden

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

    The neuron is detecting section or paragraph boundaries by firing on the very first word (often a capitalized token) at the start of a new text block.

    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
     marít
    0.86
     Wärm
    0.80
     Jiménez
    0.80
     ⋅
    0.80
     Guz
    0.79
     dure
    0.79
     hemod
    0.78
     mehndi
    0.77
     Ừ
    0.77
     tarn
    0.76
    POSITIVE LOGITS
    сколько
    0.80
    গ
    0.79
    gl
    0.79
    ти
    0.76
    ер
    0.76
    ху
    0.74
    Те
    0.73
    s
    0.73
    সূত্র
    0.71
    ளில்
    0.70
    Activations Density 0.000%

    No Known Activations