© 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. 3532
    Prev
    Next
    INDEX
    Explanations

    This neuron fires on the first content word at the beginning of a new sentence or block (i.e. the word immediately following a sentence/paragraph break).

    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
    ana
    0.65
    ining
    0.64
    aaa
    0.64
    =
    0.62
     L
    0.61
    ov
    0.61
    iy
    0.59
    inis
    0.59
     McManus
    0.59
    yla
    0.59
    POSITIVE LOGITS
    ْت
    0.72
    cento
    0.72
    annuation
    0.72
     mettant
    0.70
    cence
    0.70
    gence
    0.69
    plätze
    0.66
    fait
    0.65
     événement
    0.65
    dokter
    0.64
    Activations Density 0.945%

    No Known Activations