© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 68817
    Prev
    Next
    INDEX
    Explanations

    Proper nouns preceded by periods

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

    This neuron fires most strongly on words that are part of section or article titles and headings—i.e. tokens appearing in document titles.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    cambrian
    -0.88
    řízení
    -0.85
     dvě
    -0.82
     zprávy
    -0.81
    ürtel
    -0.80
    ))^
    -0.79
    bedaan
    -0.79
    imbangan
    -0.79
    isition
    -0.78
    ándo
    -0.78
    POSITIVE LOGITS
     kei
    0.90
    ah
    0.89
    wa
    0.85
    orton
    0.85
    gangss
    0.84
    mmung
    0.84
    ény
    0.83
    ru
    0.83
     هنگ
    0.83
    ��������
    0.82
    Activations Density 0.496%

    No Known Activations