© 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. 2113
    Prev
    Next
    INDEX
    Explanations

    The neuron flags the very first content word of each new text segment (i.e. the initial token right after a <start> tag).

    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
    ir
    0.80
    न
    0.78
     Tracks
    0.75
    </em>
    0.74
    ka
    0.74
    th
    0.73
    kar
    0.70
    in
    0.70
    '
    0.69
    Y
    0.69
    POSITIVE LOGITS
     steadily
    0.86
    sadpoetry
    0.86
     reagent
    0.85
    lcnaf
    0.84
    ﮯ
    0.83
    LOTREntity
    0.82
     ಈಗ
    0.81
     abscess
    0.81
     dimensionless
    0.81
     radionuclide
    0.79
    Activations Density 0.000%

    No Known Activations