© 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. 11421
    Prev
    Next
    INDEX
    Explanations

    explaining topics and ideas

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

    The neuron strongly activates on tokens that begin new paragraphs or segments (i.e. paragraph‐initial words).

    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
     पर
    0.75
     la
    0.74
     عبد
    0.74
     del
    0.73
    ور
    0.73
    א
    0.72
     manager
    0.71
     風
    0.71
    なく
    0.71
     da
    0.70
    POSITIVE LOGITS
    emper
    0.89
    क्शन
    0.84
    yfikacja
    0.82
    ম্বা
    0.81
    лены
    0.80
    transferases
    0.80
    ების
    0.78
    ხილ
    0.76
    yong
    0.75
    伱
    0.75
    Activations Density 0.001%

    No Known Activations