© 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. 10-GEMMASCOPE-RES-131K
    7. 105772
    Prev
    Next
    INDEX
    Explanations

    Bureau of Alcohol, Tobacco, Firearms

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

    The neuron fires on tokens that appear in titles, headlines, or other section headings (including both title-cased content words and the little function words “the” and “and” when they’re part of a title).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     позволяет
    -1.42
     любых
    -1.32
    ляет
    -1.27
     sogar
    -1.27
     показывает
    -1.23
     включает
    -1.21
     Robles
    -1.20
     ”
    -1.17
     to
    -1.17
     Allows
    -1.16
    POSITIVE LOGITS
    Ţ
    1.33
     which
    1.30
     الخم
    1.29
    饜
    1.26
     moest
    1.23
     convencer
    1.21
     negociar
    1.20
     TAMBÉM
    1.18
     台中
    1.18
     this
    1.18
    Activations Density 0.059%

    No Known Activations