© 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. 31-GEMMASCOPE-2-RES-16K
    4. 29
    Prev
    Next
    INDEX
    Explanations

    here's

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

    The neuron strongly activates on the word “There,” especially when it begins an existential construction (e.g. “There is/are…”).

    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_31_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
    これにより
    0.31
    みて
    0.30
    さて
    0.29
    そのため
    0.29
    เมื่อ
    0.29
    その後
    0.29
    すと
    0.29
     %>.
    0.28
    見て
    0.28
     চালিয়ে
    0.28
    POSITIVE LOGITS
     are
    0.88
     aren
    0.76
     są
    0.68
     were
    0.67
     हैं
    0.66
     is
    0.65
     isn
    0.63
     exists
    0.63
     janë
    0.63
     आहेत
    0.62
    Activations Density 0.238%

    No Known Activations