© 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. 6614
    Prev
    Next
    INDEX
    Explanations

    greeting or start of phrase

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

    This neuron is sensitive to numeric metadata tokens (the floating-point numbers) rather than to actual words.

    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
     dicen
    0.67
     ->
    0.66
     daca
    0.66
     Straf
    0.66
     +.
    0.65
     +
    0.64
     geek
    0.63
     nyc
    0.62
     coment
    0.62
     ++
    0.61
    POSITIVE LOGITS
    Hello
    0.83
    Olá
    0.83
    Здравствуйте
    0.79
    Dear
    0.73
    Congratulations
    0.73
    Merhaba
    0.73
    Bonjour
    0.73
    Добрый
    0.72
    Hallo
    0.70
    Ві
    0.70
    Activations Density 0.230%

    No Known Activations