© 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. 3837
    Prev
    Next
    INDEX
    Explanations

    instructional requests

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

    The neuron is firing on words that start a new sentence or segment—i.e. the very first content token (often a capitalized word) at the beginning of a sentence or after a strong 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
     Izv
    0.88
    ienka
    0.83
    iesi
    0.82
     نهايه
    0.79
    ouwd
    0.78
    lbrace
    0.78
    alade
    0.77
    礙
    0.77
    geq
    0.77
    ىسى
    0.75
    POSITIVE LOGITS
     needless
    0.70
     smelling
    0.70
     metros
    0.67
     tuto
    0.67
     todas
    0.67
     hours
    0.66
     Biden
    0.66
    rá
    0.66
     Eric
    0.66
     Nicholson
    0.66
    Activations Density 0.000%

    No Known Activations