© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 13063
    Prev
    Next
    INDEX
    Explanations

    numbers and multiplication

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

    The neuron activates on numeric tokens—digits and measurements such as counts, page numbers, quantities, and dates.

    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
    表達
    0.57
    cache
    0.54
    anchor
    0.52
    дам
    0.52
    Liên
    0.50
    Orient
    0.50
    Erd
    0.49
    桷
    0.49
    葬
    0.49
    Producto
    0.49
    POSITIVE LOGITS
     imperative
    0.51
     subsystem
    0.50
    nje
    0.50
    ška
    0.48
     Bernal
    0.48
     laying
    0.47
     grips
    0.47
    ilin
    0.47
    alian
    0.47
     Lehman
    0.46
    Activations Density 0.079%

    No Known Activations