© 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. 19751
    Prev
    Next
    INDEX
    Explanations

    a noun

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

    The neuron most strongly activates on the indefinite article “a” (and to a lesser extent “an”) when it appears in technical or patent‐style descriptions.

    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
     many
    -1.44
     medži
    -1.38
     things
    -1.38
    complished
    -1.34
     new
    -1.34
    を作り
    -1.32
    acją
    -1.32
     hundreds
    -1.32
    astricht
    -1.27
    新
    -1.27
    POSITIVE LOGITS
     và
    1.73
    それを
    1.60
    vieto
    1.59
    紲
    1.52
    それに
    1.48
     경우
    1.42
    上記の
    1.42
    中略
    1.41
    t
    1.40
     и
    1.39
    Activations Density 0.320%

    No Known Activations