© 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 EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 21389
    Prev
    Next
    INDEX
    Explanations

    Test or lab contexts

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

    The neuron fires on short technical labels and acronyms—things like test names (“test‐dropdown,” Unit, Fact), study‐design terms (“single‐center”), program codes (CAED, ATD), and genus names (Vib­rio)—i.e. compact domain-specific keywords or identifiers.

    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
     on
    -2.42
     by
    -2.30
     in
    -2.14
     or
    -2.13
     when
    -2.03
     "
    -1.95
     into
    -1.86
     and
    -1.83
     for
    -1.82
     to
    -1.81
    POSITIVE LOGITS
    錾
    1.73
    ISCHE
    1.70
     estadounid
    1.68
     pourront
    1.67
    ٨
    1.66
     illustre
    1.65
     masaj
    1.64
     Biblia
    1.63
    
    1.63
     each
    1.58
    Activations Density 0.168%

    No Known Activations