© 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. 31-GEMMASCOPE-2-RES-16K
    7. 4936
    Prev
    Next
    INDEX
    Explanations

    content warnings, feature gate, analysis paralysis

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

    The neuron strongly activates on single-word nouns that denote key concepts or topics (e.g. “step,” “point,” “mpeg,” “track,” “market,” etc.).

    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_31_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
     Reality
    0.64
    ö
    0.62
    ↵
    0.61
     Social
    0.61
    ㇂
    0.61
     Red
    0.60
     Example
    0.60
     Info
    0.59
    ar
    0.58
    ̿
    0.58
    POSITIVE LOGITS
     waarmee
    0.70
    版本的
    0.70
     لجنة
    0.69
    உலகின்
    0.69
     tirelessly
    0.67
    iguing
    0.64
     embroiled
    0.64
    ர்ந்து
    0.63
     siitä
    0.63
    ographed
    0.63
    Activations Density 0.958%

    No Known Activations