© 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. 24-GEMMASCOPE-2-RES-16K
    7. 8436
    Prev
    Next
    INDEX
    Explanations

    The neuron fires strongly on isolated, content-bearing “keyword” tokens (i.e. comparatively rare, informative words) rather than on common function words.

    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
    Ethnic
    0.66
    旄
    0.61
    誣
    0.60
     racial
    0.60
    녀
    0.59
    ethnic
    0.59
    ennzeichnet
    0.58
    譏
    0.57
     Racial
    0.57
    溆
    0.56
    POSITIVE LOGITS
     permette
    0.82
     hatchback
    0.80
     ergonomics
    0.78
     ergonom
    0.76
     Garmin
    0.75
     does
    0.74
    #,
    0.74
     Revit
    0.74
    verting
    0.73
     software
    0.72
    Activations Density 0.000%

    No Known Activations