© 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. 92953
    Prev
    Next
    INDEX
    Explanations

    programming tools and versions

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

    The neuron strongly activates on programming‐style tokens—things like numeric literals, version strings, macro or constant identifiers, file/include names, and other code identifiers—rather than ordinary prose.

    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
    Breaker
    -0.79
    freunde
    -0.78
     pí
    -0.77
     Ausführung
    -0.77
    Läh
    -0.76
     compradores
    -0.75
     Sehingga
    -0.75
    olucion
    -0.74
    AuxInt
    -0.74
     Bentuk
    -0.74
    POSITIVE LOGITS
     inappropriés
    0.95
     vœ
    0.87
     toilette
    0.85
    LOYD
    0.82
     mostrarán
    0.77
    iegler
    0.75
     Emo
    0.74
    getKeyCode
    0.73
    函数
    0.73
     pflegen
    0.72
    Activations Density 0.045%

    No Known Activations