© 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. 2014
    Prev
    Next
    INDEX
    Explanations

    Middle Asia

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

    The neuron fires strongly on multi‐word names and titles or other standout compounds—especially hyphenated terms, repeated words, and all-caps or title-style phrases—i.e. proper names and named entities.

    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
    RMSE
    0.59
    Eppo
    0.59
    zoomSeekBar
    0.58
    мат
    0.56
    مة
    0.56
    𝜋
    0.56
    дова
    0.55
    spacePad
    0.55
    hattam
    0.55
    اا
    0.54
    POSITIVE LOGITS
    -
    0.64
    )
    0.56
     [
    0.55
    :
    0.53
    .
    0.53
     :
    0.52
     https
    0.52
     successfully
    0.51
     following
    0.51
     -
    0.50
    Activations Density 0.000%

    No Known Activations