© 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. 1823
    Prev
    Next
    INDEX
    Explanations

    hexadecimal codes or identifiers

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

    This neuron is sharply tuned to numeric and code‐style tokens—such as numbers, datestamps, file names/extensions, and punctuation common in URLs or paths.

    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
     🎉
    0.71
     मित्रा
    0.69
     😄
    0.65
     🌱
    0.64
     exemplified
    0.64
     informatics
    0.63
     обязательно
    0.62
     pollutant
    0.61
     adsorbent
    0.59
     ܀
    0.59
    POSITIVE LOGITS
    0
    0.57
    改
    0.52
    6
    0.49
    %
    0.49
    ffffff
    0.48
    ッ
    0.47
    Contrast
    0.46
    CGFloat
    0.46
    NaN
    0.46
    Unknown
    0.45
    Activations Density 0.205%

    No Known Activations