© 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. 4577
    Prev
    Next
    INDEX
    Explanations

    myopic evening, theaters, function, miles, shorther

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

    The neuron strongly activates on numeric quantities—especially decimal numbers—so it’s detecting tokens that represent numerical measurements or figures.

    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
     scoperta
    0.85
    İM
    0.84
     apesar
    0.81
    ﻴ
    0.80
    ti
    0.79
    eers
    0.79
     descoberta
    0.79
     Apesar
    0.78
    ؘ
    0.78
     comunicação
    0.77
    POSITIVE LOGITS
    ث
    0.88
    Т
    0.85
     democrat
    0.84
    स
    0.81
    ص
    0.80
    с
    0.77
    T
    0.77
    خ
    0.77
    ول
    0.76
    رو
    0.76
    Activations Density 0.000%

    No Known Activations