© 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. 15519
    Prev
    Next
    INDEX
    Explanations

    `Key` and punctuation

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

    This neuron fires strongly on the very first content‐bearing token of each text segment, i.e. it’s detecting the start of a new block of text.

    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
     соответствующих
    0.89
     специализирован
    0.87
    Jr
    0.84
     рассматри
    0.82
     миллиар
    0.82
    "./
    0.82
    "
    0.82
    $;
    0.81
     соответству
    0.80
    ")],
    0.79
    POSITIVE LOGITS
     dunque
    0.79
    ក្រោយ
    0.75
    記者
    0.75
     diritto
    0.73
    безпе
    0.72
    ್ಣ
    0.71
    记者
    0.70
    зів
    0.70
     azioni
    0.70
    議
    0.70
    Activations Density 0.001%

    No Known Activations