© 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
    2. Gemma-2-27B
    3. 22-GEMMASCOPE-RES-131K
    4. 84072
    Prev
    Next
    INDEX
    Explanations

    program structure tags

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

    This neuron detects metadata or “header” tokens—i.e. section‐ or tag‐names and named identifiers such as documentation headings, acronyms, and code annotations (e.g. “usage,” “PRWEB,” “Reuters,” “XmlAttribute,” “cassert”).

    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
     historia
    -1.21
    世帯
    -1.05
    mètres
    -1.05
    ômetro
    -1.02
     vinos
    -1.01
     sonore
    -1.00
    xmax
    -0.99
    akteur
    -0.99
    ٿ
    -0.99
    miste
    -0.98
    POSITIVE LOGITS
    凄く
    1.40
    マジで
    1.39
     größe
    1.34
    ビックリ
    1.34
     überprüft
    1.32
     and
    1.31
     provoquer
    1.30
    為に
    1.28
    んですが
    1.27
     geprüft
    1.25
    Activations Density 0.002%

    No Known Activations