© 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. 7586
    Prev
    Next
    INDEX
    Explanations

    conditional logic words

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

    This neuron fires strongly at sentence boundaries—i.e. the period at the end of a sentence and the capitalized token starting the next sentence.

    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
    Một
    0.61
    ᱓
    0.60
     ).
    0.58
     ）
    0.55
    ।]
    0.55
     Vào
    0.55
    .).
    0.55
     Durchführung
    0.54
     Telugu
    0.54
    ）
    0.54
    POSITIVE LOGITS
    resistant
    0.57
     surface
    0.53
    claiming
    0.50
     nons
    0.50
    raum
    0.50
    surface
    0.50
     parliamentary
    0.48
    기관
    0.48
     dams
    0.48
    usable
    0.47
    Activations Density 0.000%

    No Known Activations