© 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. 7139
    Prev
    Next
    INDEX
    Explanations

    vector direction calculations

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

    The neuron strongly lights up on the very first meaningful content word (the first non‐trivial token) after each paragraph or section break.

    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
     überhaupt
    0.78
    كية
    0.76
     Pemerintah
    0.75
     ebenso
    0.75
    ֔
    0.73
     Shaikh
    0.73
    ії
    0.73
     Aś
    0.73
    ômes
    0.71
    पृथ
    0.71
    POSITIVE LOGITS
     이때
    0.72
    F
    0.71
    README
    0.70
     ভাইকে
    0.68
    url
    0.67
     thumbnail
    0.67
    V
    0.67
    その
    0.67
    出一个
    0.67
    上限
    0.66
    Activations Density 0.000%

    No Known Activations