© 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. 296
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly fires on quantitative or specification language—especially numbers, measurements, units, and “according to/depending on” style parameter phrases.

    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
     Abbiamo
    0.40
     Một
    0.40
     を
    0.39
     Announces
    0.38
    스를
    0.38
     цьому
    0.37
     لإ
    0.37
    を満
    0.36
     Nucleaire
    0.36
    َع
    0.35
    POSITIVE LOGITS
     location
    0.99
     size
    0.93
     characteristics
    0.89
     geography
    0.86
     ধরন
    0.84
     types
    0.83
     type
    0.83
    location
    0.83
     sizes
    0.82
    类型
    0.81
    Activations Density 6.126%

    No Known Activations