© 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. 95596
    Prev
    Next
    INDEX
    Explanations

    citations like [@cit] or [cit#]

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

    The neuron consistently fires on specialized domain or technical identifiers—like mathbb, Razor, Movie, Context, cursor, processing, field—i.e. tokens naming technical concepts, classes, modules, or mathematical structures.

    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
     biela
    -1.48
     工房
    -1.37
     bojo
    -1.35
     火鍋
    -1.32
     déli
    -1.32
    綃
    -1.30
     artesan
    -1.23
     coute
    -1.23
     gelatina
    -1.22
     kabát
    -1.22
    POSITIVE LOGITS
    u
    1.41
    p
    1.40
    3
    1.30
    h
    1.30
    5
    1.27
    7
    1.27
    f
    1.26
    w
    1.25
    on
    1.23
    r
    1.23
    Activations Density 0.061%

    No Known Activations