© 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. 13292
    Prev
    Next
    INDEX
    Explanations

    radar reflections

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

    The neuron is tuned to detect numeric expressions—especially technical or specification numbers (e.g. version numbers, measurements, and decimal values).

    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
     photoshop
    0.63
     actomyosin
    0.63
     lightbox
    0.63
    鸬
    0.62
     neurotic
    0.61
     poissons
    0.60
     Photoshop
    0.59
     pipette
    0.59
    堂
    0.57
     rocker
    0.57
    POSITIVE LOGITS
     antenna
    1.86
     antennas
    1.82
     Antenna
    1.79
     антен
    1.69
     antena
    1.66
    antenna
    1.65
     radar
    1.53
     VHF
    1.49
     antennes
    1.46
     Microwave
    1.45
    Activations Density 0.244%

    No Known Activations