© 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. 31-GEMMASCOPE-2-RES-16K
    7. 1221
    Prev
    Next
    INDEX
    Explanations

    A followed by a word

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

    This neuron strongly activates on the first content word of a new paragraph or section (i.e. sentence-initial 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_31_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
    "),
    0.59
     ⓒ
    0.54
     ausschließlich
    0.54
    <unused2164>
    0.54
     Interiors
    0.54
     ($\
    0.53
     אך
    0.53
    <unused2199>
    0.53
    }")
    0.53
     Orth
    0.52
    POSITIVE LOGITS
    cknowled
    1.05
    cknow
    0.96
    verages
    0.95
    fonso
    0.81
    romatic
    0.80
    yrıca
    0.79
    ileen
    0.74
    rugula
    0.73
    ffiche
    0.73
     menudo
    0.71
    Activations Density 0.376%

    No Known Activations