© 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. 10043
    Prev
    Next
    INDEX
    Explanations

    detour, camouflage, monogram, SUPREME

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

    The neuron strongly responds to words denoting diversion or rerouting—terms like “detour,” “rerouting,” “camouflage,” etc., that imply taking an alternate path or hiding.

    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
    냑
    -2.55
    妧
    -2.31
    лением
    -2.19
    
    -2.19
    읭
    -2.17
    ܃
    -2.16
    蟊
    -2.14
    쨩
    -2.13
    美国的
    -2.09
     実例
    -2.08
    POSITIVE LOGITS
    ↵
    2.86
     Other
    2.19
    譬
    2.08
    ,”
    2.06
     
    1.98
    把你
    1.93
     “
    1.91
     berbagai
    1.90
    とその
    1.90
     was
    1.87
    Activations Density 0.001%

    No Known Activations