© 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. 31-GEMMASCOPE-2-RES-16K
    7. 10542
    Prev
    Next
    INDEX
    Explanations

    millions of and trillions of

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

    The neuron primarily detects the word “millions” (as in “millions of …”), flagging expressions of very large counts.

    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
     অঙ্গ
    0.60
     প্রক্রিয়া
    0.58
     notes
    0.56
     essentiel
    0.55
     elabor
    0.54
     sostanz
    0.54
    notes
    0.53
     /><
    0.52
     folders
    0.50
    ঈশ্বর
    0.49
    POSITIVE LOGITS
    初の
    0.76
     ব্যবহৃত
    0.61
     নিয়েছে
    0.60
     распоря
    0.58
    संह
    0.57
    စား
    0.57
    扱う
    0.56
     ერთი
    0.56
     হয়েছে
    0.55
     recuperar
    0.55
    Activations Density 0.001%

    No Known Activations