© 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. 75215
    Prev
    Next
    INDEX
    Explanations

    . relations agencies

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

    The neuron strongly activates on all-caps sequences, single-letter initials, and uppercase abbreviations (with or without trailing periods), effectively detecting acronyms and similar uppercase tokens.

    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
     enhances
    -1.32
     first
    -1.30
     essential
    -1.27
     routinely
    -1.27
     enticing
    -1.24
     multifaceted
    -1.24
     increases
    -1.21
     foolish
    -1.21
     tailored
    -1.20
    おわりに
    -1.19
    POSITIVE LOGITS
     xviii
    1.27
     饼
    1.22
    phous
    1.18
    และ
    1.18
     Redmi
    1.16
     favoriser
    1.16
     quà
    1.13
    somewhat
    1.13
     Radeon
    1.13
     vetores
    1.10
    Activations Density 0.015%

    No Known Activations