© 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. 1634
    Prev
    Next
    INDEX
    Explanations

    The neuron fires strongly on tokens that begin new sections or paragraphs—i.e. it flags the first content word of a fresh block or heading.

    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
     maturities
    0.79
     své
    0.75
     mM
    0.73
     taxas
    0.71
     allons
    0.69
    ლებ
    0.68
    o
    0.68
     pude
    0.67
     exons
    0.67
     tane
    0.66
    POSITIVE LOGITS
    crab
    0.78
    гүй
    0.78
     whey
    0.75
    què
    0.75
    ᵕ
    0.72
     SCP
    0.72
    स्पून
    0.71
     Observer
    0.71
    гура
    0.70
    coast
    0.68
    Activations Density 0.161%

    No Known Activations