© 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. 11982
    Prev
    Next
    INDEX
    Explanations

    as something happens

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

    This neuron fires strongly on the very first word of a new segment or sentence, marking sentence-initial tokens.

    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
     had
    0.93
     after
    0.91
    :
    0.81
     that
    0.76
     when
    0.76
    _
    0.75
     which
    0.74
    ’
    0.71
     could
    0.70
     Had
    0.69
    POSITIVE LOGITS
    ುತ್ತವೆ
    0.80
    ymptotic
    0.75
    bardziej
    0.75
     متنوع
    0.74
    bestos
    0.73
    perity
    0.73
     progresses
    0.73
     нагре
    0.73
    भवती
    0.71
     verletzt
    0.71
    Activations Density 0.121%

    No Known Activations