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

    Neuronpedia

    INDEX
    Explanations

    how to do something effectively

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     relying
    0.49
     relies
    0.48
     dwells
    0.48
     будь
    0.46
     inventing
    0.44
    只会
    0.44
     awaits
    0.43
     consultancy
    0.42
     reliance
    0.42
     conformity
    0.41
    POSITIVE LOGITS
     effectively
    0.90
     properly
    0.88
     maximize
    0.84
     best
    0.81
     approach
    0.78
     safely
    0.78
     deal
    0.75
     proceed
    0.73
     Effectively
    0.73
     Properly
    0.72
    Act Density 0.091%

    No Known Activations