© 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. 14413
    Prev
    Next
    INDEX
    Explanations

    Questions and analyses

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

    The neuron reliably spikes on the first substantial content word after each blank‐line or section break, i.e. it marks the start of new paragraphs or document segments.

    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
    ObjectUnderTest
    0.87
    ي
    0.86
    ק
    0.84
    ig
    0.80
    ни
    0.79
     однако
    0.79
    ک
    0.79
    क
    0.78
    有望
    0.78
    selves
    0.77
    POSITIVE LOGITS
     Polynesian
    0.86
     bohemian
    0.84
     pathologist
    0.84
     deposition
    0.83
     quince
    0.82
     pà
    0.82
     chromatin
    0.80
     quinoa
    0.80
     jacuzzi
    0.80
     bisque
    0.80
    Activations Density 0.000%

    No Known Activations