© 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. 4682
    Prev
    Next
    INDEX
    Explanations

    language complexity levels

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

    The neuron fires strongly on the last token of a sequence chunk, i.e. it detects the end-of-segment boundary.

    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
     Observer
    0.88
     Tempat
    0.86
    therly
    0.85
    ىل
    0.83
     seura
    0.78
    <unused67>
    0.77
    jų
    0.77
     نيك
    0.76
     opět
    0.75
     conçu
    0.74
    POSITIVE LOGITS
    ק
    0.78
    שת
    0.75
    머
    0.71
     cari
    0.71
    СС
    0.71
    ג
    0.70
    лась
    0.67
    מר
    0.66
     tada
    0.66
    פ
    0.66
    Activations Density 0.000%

    No Known Activations