© 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. 20393
    Prev
    Next
    INDEX
    Explanations

    programming constructs

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

    The neuron fires on programming‐language tokens—identifiers, keywords, and property names—rather than ordinary natural-language words.

    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
     such
    -1.88
     even
    -1.52
     also
    -1.31
    ↵
    -1.28
     may
    -1.27
     although
    -1.26
     näin
    -1.24
     pulls
    -1.24
     as
    -1.21
     Abfall
    -1.21
    POSITIVE LOGITS
    醅
    1.60
    鸲
    1.55
    昻
    1.49
    ⎆
    1.48
    ॉप
    1.43
    訁
    1.41
     patate
    1.40
    iações
    1.38
    諗
    1.36
    ꯱
    1.34
    Activations Density 0.075%

    No Known Activations