© 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
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 3048
    Prev
    Next
    INDEX
    Explanations

    key roles and components

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

    This neuron detects fragments of compound role titles that signal joint or auxiliary positions (e.g. the “co-” in co-developer/editor/author/founder, “vice” in vice-president, and “founding” in founding member).

    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
     welches
    0.68
     upperBound
    0.68
    꼭
    0.65
    സ്
    0.63
    tywn
    0.62
     Commission
    0.58
    ullah
    0.57
    um
    0.56
     College
    0.55
     nutshell
    0.55
    POSITIVE LOGITS
     рядом
    0.73
     автором
    0.66
     גי
    0.64
     copos
    0.64
     имели
    0.63
     plotter
    0.63
     актриса
    0.63
    শিল্পী
    0.62
    Ⲣ
    0.62
     planer
    0.62
    Activations Density 0.018%

    No Known Activations