© 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. 54272
    Prev
    Next
    INDEX
    Explanations

    metaphorical or abstract concepts

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

    This neuron activates on multi-word noun compounds—especially formal names of institutions or specialized items (e.g. “Claremont Graduate University,” “Oxford University,” “fetal ultrasound machines,” “sports as a platform”).

    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
     vêtement
    -0.83
     each
    -0.81
     between
    -0.81
     skry
    -0.80
     dynasty
    -0.80
     Stell
    -0.76
    ちな
    -0.75
     altri
    -0.73
     other
    -0.73
     واب
    -0.72
    POSITIVE LOGITS
    setObject
    0.92
     profesionales
    0.87
    MessageInfo
    0.86
    UIImage
    0.86
     metodología
    0.85
    SUPPORT
    0.84
    authenticate
    0.84
    ⌀
    0.84
    erdere
    0.83
    setStatus
    0.83
    Activations Density 0.042%

    No Known Activations