© 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. 22-GEMMASCOPE-RES-131K
    4. 35989
    Prev
    Next
    INDEX
    Explanations

    Russian months and names

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

    The neuron consistently fires on standalone nouns or name‐like tokens (object/resource names) rather than function words or syntax.

    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
    se
    -1.18
    particularly
    -1.16
    sub
    -1.13
    lové
    -1.09
     myriad
    -1.09
    the
    -1.08
    Therefore
    -1.07
     certaines
    -1.06
    so
    -1.04
    ਃ
    -1.03
    POSITIVE LOGITS
     at
    1.11
     октября
    1.10
     сентября
    1.07
     декабря
    1.05
    Olga
    1.04
    Inggris
    1.02
     ноября
    1.02
    blusa
    1.01
     avenida
    1.01
     февраля
    1.00
    Activations Density 0.020%

    No Known Activations