© 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. 24-GEMMASCOPE-2-RES-16K
    7. 11603
    Prev
    Next
    INDEX
    Explanations

    numbers, punctuation, and short word fragments

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

    The neuron primarily responds to numeric tokens and number-like expressions (years, figures, percentages, etc.).

    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_24_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
     gunmen
    0.90
     Reporters
    0.88
     advisers
    0.82
     businessmen
    0.80
    ɚ
    0.79
     staffers
    0.78
     reporters
    0.77
     Economist
    0.77
     lawmakers
    0.77
     businessman
    0.76
    POSITIVE LOGITS
    最後まで
    0.62
    に係る
    0.61
    DEFAULT
    0.57
     (!(
    0.56
    のもの
    0.56
    ()));
    0.56
     pertaining
    0.55
    ();
    0.55
     aquello
    0.55
     thereof
    0.54
    Activations Density 0.005%

    No Known Activations