© 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. 10178
    Prev
    Next
    INDEX
    Explanations

    task instructions

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

    The neuron fires on standalone section‐label tokens (things like “A:”, “Notes:”, “Description:”, “This wikiHow teaches…”, etc.) that mark the start of a new heading or instructional segment.

    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
     prospective
    0.70
     futurs
    0.68
    ="
    0.67
     Prodi
    0.66
    _{
    0.65
     renewables
    0.64
     Mohd
    0.63
     fisi
    0.62
     müşteri
    0.62
     linseed
    0.61
    POSITIVE LOGITS
    ist
    0.63
    ซึ่ง
    0.62
    ний
    0.59
    馊
    0.59
    ติด
    0.57
     যার
    0.57
    ইন
    0.57
     어렵
    0.55
    вори
    0.54
    בית
    0.54
    Activations Density 0.000%

    No Known Activations