© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 7923
    Prev
    Next
    INDEX
    Explanations

    configuration / Partials

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

    The neuron strongly activates on programming identifiers and keywords—especially capitalized API calls, annotations, and configuration tokens (e.g. “Enable,” “Auto,” “DEFINE,” “cloud,” “Suppress”).

    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
    section
    -1.84
    ↵↵↵
    -1.71
    瓩
    -1.69
     beinahe
    -1.62
     “(
    -1.62
     находятся
    -1.55
     she
    -1.54
    
    -1.53
     prickly
    -1.53
     zuerst
    -1.53
    POSITIVE LOGITS
     zapatillas
    1.53
    Ajouter
    1.53
     вашу
    1.52
    瞜
    1.49
    芣
    1.45
    teilungen
    1.44
    怱
    1.44
     Nazionale
    1.43
    votre
    1.41
    鷽
    1.41
    Activations Density 0.006%

    No Known Activations