© 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. 104581
    Prev
    Next
    INDEX
    Explanations

    miniature safety waterproof case

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

    The neuron strongly activates on isolated alphanumeric labels—tokens made up of mostly uppercase letters, numbers, or mixed letter‐digit sequences (e.g. product codes, set sizes, 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
    :),
    -1.34
    °;
    -1.31
    +++++
    -1.27
    も多く
    -1.21
     観光
    -1.20
     garansi
    -1.20
    ⿷
    -1.18
    🫴
    -1.17
     algas
    -1.17
    tifiable
    -1.17
    POSITIVE LOGITS
    ございません
    1.22
     Each
    1.18
    A
    1.11
    M
    1.09
     As
    1.09
     Now
    1.07
     Couldn
    1.06
    Very
    1.06
    rodalom
    1.05
     somber
    1.05
    Activations Density 0.039%

    No Known Activations