© 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. 22-GEMMASCOPE-RES-131K
    7. 35716
    Prev
    Next
    INDEX
    Explanations

    specific phrases and titles

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

    The neuron is detecting mentions of “fun” and other strongly positive, enjoyment-oriented words describing pleasurable or upbeat experiences.

    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
    𝒙
    -1.45
     Küche
    -1.25
     or
    -1.20
     أشهر
    -1.17
    pushd
    -1.16
    累了
    -1.13
    iços
    -1.13
    ciri
    -1.11
    𝛽
    -1.10
     familiar
    -1.07
    POSITIVE LOGITS
     هناك
    1.61
     there
    1.50
     myself
    1.48
    dokument
    1.41
     Ketika
    1.41
     Misalnya
    1.38
    𝑈
    1.38
     Bagaimana
    1.36
     khán
    1.34
     Especific
    1.30
    Activations Density 0.003%

    No Known Activations