© 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. 36333
    Prev
    Next
    INDEX
    Explanations

    category specific attributes

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

    The neuron strongly activates on tokens that look like structured identifiers or field/key names—i.e. CamelCase or all‐caps words and labels you’d see in code, schema definitions, or table headings.

    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
     interessant
    -1.24
     kompatibel
    -1.16
     feront
    -1.14
    LLLL
    -1.10
     seront
    -1.05
     gekocht
    -1.04
     vertus
    -1.02
     pério
    -0.99
     konstant
    -0.99
     konfig
    -0.99
    POSITIVE LOGITS
     type
    1.22
    шення
    1.21
     as
    1.19
     size
    1.16
     vergüenza
    1.12
     тип
    1.10
     spodnie
    1.09
     itself
    1.09
     specific
    1.09
    ‌های
    1.08
    Activations Density 0.050%

    No Known Activations