© 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. 12716
    Prev
    Next
    INDEX
    Explanations

    hereinafter, hereafter, herein

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

    The neuron detects defined‐term declarations in legal/contract text—that is, words or phrases introduced in parentheses and quotation marks as formal definitions.

    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
    GORITH
    0.62
     MANY
    0.61
    😦
    0.61
    MANY
    0.60
    👳
    0.59
     FIXME
    0.58
    Many
    0.58
    BEGIN
    0.58
     Whose
    0.57
     {/*
    0.55
    POSITIVE LOGITS
    hereinafter
    0.80
    hereafter
    0.72
     hereinafter
    0.71
     hereafter
    0.66
     collett
    0.65
    collect
    0.64
     collective
    0.61
     colectivo
    0.58
    这里的
    0.56
    蚝
    0.56
    Activations Density 0.018%

    No Known Activations