© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 90873
    Prev
    Next
    INDEX
    Explanations

    initialisms with numbers or Roman numerals

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

    The neuron selectively activates on alphanumeric technical identifiers—uppercase names, acronyms, and model‐number–style tokens (e.g. Drosophila, Radeon HD, IFN‐α, iPhone 6).

    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
     because
    -1.83
     его
    -1.52
    ถ้า
    -1.47
    了一会儿
    -1.47
     usually
    -1.46
    しかも
    -1.45
     classy
    -1.45
     also
    -1.43
     Usually
    -1.43
    他
    -1.38
    POSITIVE LOGITS
     to
    2.28
     alcune
    1.56
     of
    1.55
    いくつかの
    1.47
     desenho
    1.44
     mismas
    1.41
    mamp
    1.39
     inmiddels
    1.39
    several
    1.39
    seems
    1.38
    Activations Density 0.222%

    No Known Activations