© 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. 1813
    Prev
    Next
    INDEX
    Explanations

    pound, nudge, minigame, adultery, Russian, Catalan

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

    The neuron fires strongly on common Latin‐derived prefixes at the start of multi‐syllable technical or procedural terms.

    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
    gmail
    0.65
    ?
    0.63
     Barclays
    0.62
    N
    0.62
    0.61
    e
    0.61
    米
    0.59
     Wayback
    0.58
    u
    0.58
     State
    0.58
    POSITIVE LOGITS
    ‌های
    0.71
     КО
    0.69
     altres
    0.68
     să
    0.65
     ну
    0.65
     інших
    0.64
     HYDRO
    0.64
     של
    0.63
     της
    0.63
     НА
    0.63
    Activations Density 0.048%

    No Known Activations