© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 24-GEMMASCOPE-2-RES-16K
    7. 12437
    Prev
    Next
    INDEX
    Explanations

    colorful twist

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

    The neuron is acting as a simple paragraph‐ or segment‐boundary detector, firing once at the start of each new text block (i.e. the first “content” token of each paragraph).

    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
    era
    0.70
    الا
    0.66
     informations
    0.64
    ъ
    0.64
    anta
    0.63
    forderungen
    0.62
    内容は
    0.62
    `;
    0.61
    0.61
    esar
    0.61
    POSITIVE LOGITS
     KNN
    0.88
     pyrolysis
    0.87
     dưới
    0.84
     COPD
    0.82
    চরিত
    0.82
     EET
    0.81
    OpenGL
    0.80
     seja
    0.80
     phyll
    0.79
     furry
    0.79
    Activations Density 0.000%

    No Known Activations