© 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. 12685
    Prev
    Next
    INDEX
    Explanations

    after itself"""

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

    The neuron fires strongly on uncommon, content-bearing tokens (i.e. rarer “keyword” words rather than frequent 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_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
    /
    0.91
    -/
    0.79
    /}
    0.68
     частности
    0.65
    ฯ
    0.64
     ۶
    0.60
    たち
    0.60
     singkat
    0.60
     下さい
    0.60
    +)$
    0.59
    POSITIVE LOGITS
    ,
    0.86
     constitute
    0.82
     constitutes
    0.81
    KeepOriginal
    0.80
    ，
    0.77
    是多少
    0.76
     merupakan
    0.74
     enjoys
    0.74
    ،
    0.72
     adalah
    0.72
    Activations Density 0.289%

    No Known Activations