© 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. 84380
    Prev
    Next
    INDEX
    Explanations

    proper nouns followed by action or description

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

    The neuron strongly activates on single‐token proper names (people, places, brands)—i.e. capitalized named entities.

    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
    鷽
    -1.55
     perlas
    -1.46
     différents
    -1.42
     pulseras
    -1.42
    ються
    -1.41
    €”
    -1.41
    Nazionale
    -1.38
    閬
    -1.38
    شخص
    -1.38
    へ向
    -1.38
    POSITIVE LOGITS
     is
    1.80
    ↵↵
    1.61
    礎
    1.55
     have
    1.54
    これを
    1.45
    それを
    1.45
    ztes
    1.41
    でしょうね
    1.33
     can
    1.32
    ↵
    1.31
    Activations Density 0.117%

    No Known Activations