© 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. 10-GEMMASCOPE-RES-131K
    7. 10293
    Prev
    Next
    INDEX
    Explanations

    components and parts

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

    The neuron fires on multiword technical compound nouns (specialized jargon) such as “integrated circuit,” “tablet computer,” “transmission cable,” “linear motion guide bearing,” “electrical wall plates,” “print-circuit board,” and “drill pipe.”

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
    ↵
    -2.03
    .
    -1.83
     was
    -1.63
    Two
    -1.63
     Factbook
    -1.59
     this
    -1.55
     two
    -1.52
     =
    -1.48
    is
    -1.48
    ado
    -1.47
    POSITIVE LOGITS
     afectan
    1.89
    ënt
    1.76
     constituyen
    1.74
     comienzan
    1.73
     reciben
    1.70
     podrán
    1.67
     quedaron
    1.67
    kadang
    1.66
     misschien
    1.66
    tarta
    1.66
    Activations Density 0.956%

    No Known Activations