© 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. 41448
    Prev
    Next
    INDEX
    Explanations

    placeholders and code constructs

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

    The neuron activates on tokens that are purely numeric (often long digit strings or dotted numeric “version‐style” sequences).

    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
    氿
    -3.20
    蛱
    -2.94
     is
    -2.81
    菝
    -2.81
    AdvEx
    -2.75
    鼍
    -2.56
    坌
    -2.56
    ቷ
    -2.56
    翆
    -2.50
    聼
    -2.48
    POSITIVE LOGITS
    as
    3.02
    ed
    2.73
     Ironically
    2.56
     Есть
    2.56
    gebras
    2.53
     Surprisingly
    2.52
     Relatively
    2.45
    section
    2.38
     Careful
    2.34
    褔
    2.33
    Activations Density 0.004%

    No Known Activations