© 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. 6324
    Prev
    Next
    INDEX
    Explanations

    abuse or endanger

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

    The neuron activates on numerical tokens—phone numbers, percentages, statistics, decimals—i.e. it detects numerical 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
     betek
    0.54
     இந்தியா
    0.53
     வள
    0.50
    ック
    0.49
     mediast
    0.49
    啾
    0.49
     nurture
    0.48
    UNK
    0.48
     الطلاب
    0.48
    zigzag
    0.47
    POSITIVE LOGITS
    highly
    0.49
    Highly
    0.49
    stateMutability
    0.47
    ᴥ
    0.46
     ناقص
    0.46
    
    0.46
    madan
    0.46
    mapsto
    0.46
    ��
    0.46
    LoopBlend
    0.46
    Activations Density 0.302%

    No Known Activations