© 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. 5951
    Prev
    Next
    INDEX
    Explanations

    older houses

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

    The neuron strongly activates on numeric expressions—digits or spelled-out quantities—used to convey counts, measurements, or other quantitative details.

    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
     বন্যার
    0.54
    ActionEvent
    0.54
     cân
    0.54
    ्वल
    0.54
    滦
    0.52
    尽
    0.52
     यातायात
    0.50
     reactive
    0.50
    烫
    0.50
     भर्ती
    0.49
    POSITIVE LOGITS
     houses
    2.13
     house
    2.03
     rumah
    1.95
    房子
    1.92
     mansion
    1.84
    houses
    1.77
     dwellings
    1.77
     mansions
    1.76
     Houses
    1.74
    Houses
    1.73
    Activations Density 1.023%

    No Known Activations