© 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. 35811
    Prev
    Next
    INDEX
    Explanations

    names and places followed by context

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

    The neuron lights up on isolated “data‐like” tokens—things you’d find in an infobox or bullet list—such as place names, numbers, category labels, and item names rather than on running narrative.

    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
    فرة
    -1.13
     envol
    -1.12
     any
    -1.09
    tagez
    -1.09
     specifik
    -1.09
    имость
    -1.08
     Fußballspieler
    -1.07
     magnifique
    -1.07
     ingin
    -1.05
    lanan
    -1.05
    POSITIVE LOGITS
     is
    2.34
     are
    2.03
     was
    1.60
     were
    1.28
     will
    1.16
    は
    1.14
    の一
    1.10
     understandably
    1.10
     najczęściej
    1.09
    が
    1.09
    Activations Density 0.002%

    No Known Activations