© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 14522
    Prev
    Next
    INDEX
    Explanations

    The neuron is selectively activated by the very first token of new paragraphs or text blocks—that is, it flags paragraph‐ or section-starting tokens.

    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
    o
    0.97
    ر
    0.94
    er
    0.86
    u
    0.84
    0
    0.82
    an
    0.80
    q
    0.80
    R
    0.77
    z
    0.76
    el
    0.74
    POSITIVE LOGITS
     necessidades
    0.92
    婳
    0.87
    黼
    0.83
    quisitos
    0.81
    性は
    0.80
    bardziej
    0.79
    രിച്ചി
    0.79
    ilaian
    0.79
    陬
    0.79
    顗
    0.79
    Activations Density 0.135%

    No Known Activations