© 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
    2. Gemma-3-12B
    3. 31-GEMMASCOPE-2-RES-16K
    4. 15874
    Prev
    Next
    INDEX
    Explanations

    the neuron activates on numeric tokens, especially multi‐digit numbers or numeric identifiers.

    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
     disturbing
    0.55
     pož
    0.54
    ₩
    0.54
    Value
    0.54
    值为
    0.54
     Sementara
    0.54
    ఔ
    0.53
    Lb
    0.53
    更大的
    0.52
     值
    0.52
    POSITIVE LOGITS
     Denise
    0.57
    runoff
    0.55
    run
    0.54
    removed
    0.54
     Colby
    0.54
    ристо
    0.53
    jfif
    0.53
     Oficina
    0.53
    宅
    0.53
    pi
    0.51
    Activations Density 0.000%

    No Known Activations