© 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. 74932
    Prev
    Next
    INDEX
    Explanations

    specific entities with common continuations

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

    The neuron strongly activates on capitalized tokens—proper names, acronyms, and other named‐entity or identifier-like words (e.g. “Jedi,” “Piacenza,” “DNC,” “componentDidMount”).

    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
    蒞
    -2.59
    '
    -2.58
     seksi
    -2.55
    You
    -2.52
     estadounid
    -2.50
    鈇
    -2.50
     menyen
    -2.47
     rajut
    -2.44
     arkiv
    -2.42
     maksi
    -2.36
    POSITIVE LOGITS
    ᠂
    2.80
    跺
    2.56
     tamamen
    2.45
     distraught
    2.45
     hinder
    2.41
    偏偏
    2.36
     seus
    2.36
     infuse
    2.31
     sagte
    2.30
     mitigate
    2.22
    Activations Density 0.051%

    No Known Activations