© 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. 6371
    Prev
    Next
    INDEX
    Explanations

    GUIDs and non-English words

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

    The neuron is highly sensitive to fragments of hexadecimal or UUID-style identifiers (e.g. the runs of hex digits and separators in GUIDs).

    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
    Project
    0.73
    JsonResponse
    0.71
    Radi
    0.68
    UserInput
    0.68
    غ
    0.68
    Projects
    0.68
    Sonic
    0.66
    ያንዳ
    0.65
     Вики
    0.64
    Operations
    0.63
    POSITIVE LOGITS
     amputation
    0.69
     deities
    0.67
     disa
    0.66
     lemmas
    0.63
     affectionately
    0.63
     redeem
    0.62
     nucl
    0.62
     singularly
    0.62
     parishioners
    0.61
     u
    0.61
    Activations Density 0.001%

    No Known Activations