© 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
    3. Gemma-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 15736
    Prev
    Next
    INDEX
    Explanations

    LaTeX and HTML content

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

    The neuron fires on numeric tokens (e.g. version numbers, dimensions, or other numeric literals) in the text.

    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
     as
    -1.22
     after
    -1.20
     which
    -1.19
    bouncycastle
    -1.17
     when
    -1.17
    Также
    -1.16
     with
    -1.16
     those
    -1.15
    があり
    -1.15
    与此同时
    -1.14
    POSITIVE LOGITS
    為に
    1.23
     CONCEPT
    1.23
    ほら
    1.20
    ₩
    1.20
     ſi
    1.18
     contenta
    1.18
    ええ
    1.16
     parler
    1.16
    ragende
    1.16
     refroid
    1.16
    Activations Density 0.001%

    No Known Activations