© 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. 130620
    Prev
    Next
    INDEX
    Explanations

    The neuron lights up on content‐bearing tokens (proper nouns, pronouns, verbs, numbers, etc.) and stays off on common function words—i.e. it’s a content vs. function–word detector.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    personal names and actors

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     ejemplar
    -0.93
     récomp
    -0.93
    銹
    -0.87
    年中
    -0.83
    鏽
    -0.81
     Hanson
    -0.80
     établi
    -0.79
    Wendy
    -0.77
     sprays
    -0.75
    Nay
    -0.75
    POSITIVE LOGITS
     personally
    1.00
     niece
    0.88
     and
    0.84
    AppId
    0.83
     zl
    0.81
    proxy
    0.81
    PROXY
    0.80
    ஐ
    0.78
     スタンダード
    0.75
    dataSource
    0.74
    Activations Density 0.043%

    No Known Activations