© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 8744
    Prev
    Next
    INDEX
    Explanations

    Play, Start, Clean, less

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

    The neuron primarily detects unusually long or multi‐syllabic content words (i.e. infrequent, lengthy tokens).

    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
    prises
    0.77
    efully
    0.75
     capitalism
    0.74
    و
    0.73
    itness
    0.71
    सा
    0.71
    shank
    0.70
    उचर
    0.69
     Stamps
    0.69
    ïd
    0.68
    POSITIVE LOGITS
    +}^{
    0.80
     Alfvén
    0.74
     ү
    0.73
     sende
    0.71
     kurzer
    0.70
     парк
    0.70
     Dör
    0.70
    เอง
    0.70
     kön
    0.69
     ál
    0.68
    Activations Density 0.000%

    No Known Activations