© 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. 31-GEMMASCOPE-2-RES-16K
    7. 5667
    Prev
    Next
    INDEX
    Explanations

    If the

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

    This neuron strongly activates on discourse markers and conjunctions (e.g. “If,” “But,” “When,” “Yet,” “Therefore”) that introduce or link clauses.

    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
     you
    0.74
    wh
    0.68
    ഡ്
    0.66
    ots
    0.64
    抖
    0.64
    ru
    0.63
     Mom
    0.63
    í
    0.63
    il
    0.62
    šn
    0.62
    POSITIVE LOGITS
    このような
    0.81
    やっぱり
    0.75
    ;</
    0.74
     till
    0.73
     раньше
    0.73
     ইতিমধ্যে
    0.72
     هنوز
    0.71
    ,</
    0.70
     Ainda
    0.70
     στην
    0.70
    Activations Density 0.175%

    No Known Activations