© 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. 24-GEMMASCOPE-2-RES-16K
    7. 1342
    Prev
    Next
    INDEX
    Explanations

    The neuron strongly activates on numeric and date‐related tokens—e.g. years, retrieval dates, statistics—i.e. it picks out numbers and date metadata.

    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_24_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
    fono
    0.74
    뻗
    0.73
    គ្រ
    0.72
    нун
    0.70
     кеңсесинде
    0.69
     ө
    0.69
     fá
    0.68
    profissional
    0.67
    lasses
    0.67
     هذ
    0.67
    POSITIVE LOGITS
    ओम
    0.80
    空的
    0.80
    ام
    0.75
    Sec
    0.69
    い
    0.68
     towards
    0.68
    国道
    0.68
    2
    0.68
    哥哥
    0.68
    个月
    0.67
    Activations Density 0.116%

    No Known Activations