© 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. 99640
    Prev
    Next
    INDEX
    Explanations

    из

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

    The neuron fires on the Russian preposition “из” when it appears immediately before a large number (e.g. in “из 392”), i.e. part of a “… of N” page‐count or item‐count pattern.

    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
     
    -3.42
    ↵
    -3.41
      
    -2.94
     we
    -2.84
    ],
    -2.81
    "
    -2.81
     heisst
    -2.72
    ?),
    -2.70
    ↵↵
    -2.69
    </td>
    -2.66
    POSITIVE LOGITS
    鯓
    3.05
    噚
    2.81
    焿
    2.78
     pauv
    2.70
    卨
    2.64
    aktivi
    2.64
     水彩
    2.59
    蕯
    2.59
    壜
    2.53
    彣
    2.53
    Activations Density 0.001%

    No Known Activations