© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 10451
    Prev
    Next
    INDEX
    Explanations

    struct, Qt, Belief, Regulation, clay, choose

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

    The neuron is especially sensitive to technical or domain‐specific terms (e.g. “equipment,” “threshold,” “criteria”) and other formal connective words (like “and,” “will,” “all”) that commonly appear in instructional or technical listings.

    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
    ied
    0.74
     কমান্ড
    0.71
     ঘটনা
    0.70
     નથી
    0.70
    usern
    0.69
     চাহিয়া
    0.68
    arct
    0.68
    augment
    0.68
     збіль
    0.68
    centro
    0.68
    POSITIVE LOGITS
     stif
    0.78
    מ
    0.72
     maha
    0.70
    ﾗ
    0.69
     portes
    0.69
     các
    0.69
    ୀ
    0.69
     charmed
    0.68
    РО
    0.68
     hoon
    0.68
    Activations Density 0.001%

    No Known Activations