© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 123481
    Prev
    Next
    INDEX
    Explanations

    Pratt, Brennan, Gallagher, McGrath

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

    The neuron strongly activates on tokens that are the final subwords of personal names (i.e. surnames).

    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
    卝
    -3.19
    actéristi
    -3.17
    ⏑
    -3.08
     magnificent
    -3.03
     strikingly
    -3.03
     astonishing
    -3.02
     picturesque
    -2.98
     fantastic
    -2.95
     alluring
    -2.95
     exceptional
    -2.92
    POSITIVE LOGITS
    葙
    3.06
    “
    2.59
    Even
    2.55
    s
    2.45
    Being
    2.34
     sebaik
    2.34
    Not
    2.27
    Surprisingly
    2.23
    when
    2.22
    ↵
    2.20
    Activations Density 0.016%

    No Known Activations