© 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. 6448
    Prev
    Next
    INDEX
    Explanations

    The neuron is primarily activated by numerical tokens or numeric expressions in the text.

    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
    ನ
    0.76
    衩
    0.70
     ung
    0.69
     μεγαλύτε
    0.67
    а
    0.67
     εφαρμο
    0.66
    ان
    0.65
    我相信
    0.65
     document
    0.64
     flavor
    0.64
    POSITIVE LOGITS
    others
    0.82
    hashCode
    0.80
    gina
    0.80
    ௭
    0.79
    handles
    0.78
    sparse
    0.75
    balanced
    0.73
    Sparse
    0.72
    Pooled
    0.72
    packed
    0.71
    Activations Density 0.000%

    No Known Activations