© 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. 2541
    Prev
    Next
    INDEX
    Explanations

    following weather or location

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

    The neuron is firing on the very first word of each new paragraph or block (i.e. the token immediately following a blank‐line or paragraph break).

    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
    an
    0.95
    ѕ
    0.74
     Amer
    0.73
    یر
    0.73
     ORD
    0.72
    anym
    0.71
    revision
    0.70
    the
    0.69
     А
    0.68
     American
    0.68
    POSITIVE LOGITS
    auw
    0.84
    คือ
    0.78
    ancji
    0.78
     directa
    0.77
    に
    0.77
     csak
    0.75
     nét
    0.74
     diffus
    0.71
     gelijk
    0.71
     dadurch
    0.70
    Activations Density 0.001%

    No Known Activations