© 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. 12824
    Prev
    Next
    INDEX
    Explanations

    starts at, attempts to

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

    The neuron is primarily sensitive to the start‐of‐segment marker (<start>), strongly activating right at the beginning of each new text chunk.

    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
    LS
    0.86
    ্স
    0.82
     noted
    0.80
    IN
    0.79
    _{\
    0.78
    ro
    0.77
    AK
    0.77
     ভেতর
    0.77
    Blur
    0.76
    요
    0.74
    POSITIVE LOGITS
    чный
    0.86
     adjournment
    0.82
     omission
    0.81
     electors
    0.81
     basilica
    0.81
    ία
    0.80
    ಯೋಗ
    0.80
     steppe
    0.79
     pliers
    0.79
     billboards
    0.79
    Activations Density 0.000%

    No Known Activations