© 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. 5868
    Prev
    Next
    INDEX
    Explanations

    exports and Western

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

    The neuron fires most strongly on the first word of a new sentence (i.e. a capitalized, sentence-initial token).

    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
     聞い
    0.88
     遊ん
    0.87
     honti
    0.86
     Объ
    0.84
     Вин
    0.80
    格局
    0.78
     勝
    0.78
    ুগত্য
    0.77
     micelles
    0.76
     cathodes
    0.75
    POSITIVE LOGITS
    '||
    0.82
    '
    0.82
    proximity
    0.79
    s
    0.78
    Di
    0.77
     tia
    0.77
     e
    0.76
     l
    0.75
    '+
    0.75
     i
    0.74
    Activations Density 0.000%

    No Known Activations