© 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. 12881
    Prev
    Next
    INDEX
    Explanations

    structured correspondence

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

    This neuron strongly activates on the first substantive token after a sentence or clause boundary (e.g. the first content word following punctuation like “.”, “;”, “–”, or parentheses).

    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
    UNO
    0.74
     концу
    0.72
    بعة
    0.72
    ऑक्साइड
    0.71
    ớm
    0.70
    ё
    0.68
     Ontology
    0.67
     Chapters
    0.66
     CaCO
    0.66
     Chakraborty
    0.66
    POSITIVE LOGITS
    e
    0.80
     सौंद
    0.73
    indicating
    0.71
    le
    0.71
    ر
    0.71
    روس
    0.70
    ار
    0.69
    imgur
    0.69
    साठी
    0.68
    estimating
    0.68
    Activations Density 0.000%

    No Known Activations