© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 10322
    Prev
    Next
    INDEX
    Explanations

    restores, removes, generates, identifies

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

    The neuron consistently fires on content‐heavy action words—especially verbs or nominalized verbs (e.g. “treatments,” “detecting,” “building,” “fixing,” “slaughter,” “consume,” “purchase”)—highlighting phrases that describe doing or causing something.

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
     stoichiometry
    0.66
     empate
    0.66
     Governance
    0.66
    在中
    0.66
     IntelliJ
    0.64
     grd
    0.63
    🈷
    0.63
     timestep
    0.63
     ScriptInterface
    0.62
     Matemática
    0.61
    POSITIVE LOGITS
    これらの
    0.94
     these
    0.86
     이러한
    0.85
    these
    0.80
     এসব
    0.78
     انواع
    0.77
     цих
    0.73
    這些
    0.72
    这些
    0.71
     هذه
    0.70
    Activations Density 0.165%

    No Known Activations