© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 14509
    Prev
    Next
    INDEX
    Explanations

    throughout the

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

    The neuron activates on prepositional words in summary-style phrases—most prominently “of” in “highlights of…” and “over” in “over the last…,” marking out summary or wrap-up segments.

    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
    ри
    0.82
    ר
    0.74
     necesita
    0.73
    att
    0.73
    пъл
    0.71
    что
    0.71
    š
    0.71
    товар
    0.70
    ের
    0.69
    aste
    0.69
    POSITIVE LOGITS
     Throughout
    1.12
     throughout
    1.11
     במהלך
    1.08
    Throughout
    1.02
    கடந்த
    1.02
     весь
    0.98
     boyunca
    0.95
     Verlauf
    0.95
     gesamte
    0.94
    ตลอด
    0.93
    Activations Density 0.583%

    No Known Activations