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

    Neuronpedia

    INDEX
    Explanations

    code readability and expressiveness

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    thoughts
    0.66
    Think
    0.61
     shrugged
    0.61
    think
    0.59
     complej
    0.57
    Thinking
    0.55
     رابطه
    0.55
    Scheduling
    0.55
    相對
    0.54
     shrug
    0.54
    POSITIVE LOGITS
     very
    0.81
     understandable
    0.80
     easy
    0.76
     readable
    0.75
     explicit
    0.73
     quite
    0.73
     molto
    0.71
     human
    0.70
     cleaner
    0.68
     sehr
    0.67
    Act Density 0.440%

    No Known Activations