Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    INDEX
    Explanations

    helpful AI assistant refusals

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     callSettings
    0.42
     MaterialApp
    0.38
     freshest
    0.38
    乘以
    0.38
     말미암아
    0.38
    的我
    0.37
     morally
    0.36
     births
    0.36
     antérieur
    0.36
     antigenic
    0.36
    POSITIVE LOGITS
     chatbot
    0.63
    chatbot
    0.57
     assistant
    0.52
     AI
    0.49
     chatbots
    0.49
     guide
    0.48
     безопас
    0.48
    AI
    0.46
     helpful
    0.46
     innocuous
    0.46
    Act Density 0.021%

    No Known Activations