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

    Neuronpedia

    INDEX
    Explanations

    complex behaviors

    np_max-act-logits · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    <unused2164>
    0.53
    <unused2121>
    0.48
    <unused2130>
    0.48
                             
    0.46
                           
    0.45
    <unused2117>
    0.45
                            
    0.43
     Components
    0.43
                              
    0.42
     Address
    0.42
    POSITIVE LOGITS
     shenanigans
    1.08
     paranoia
    1.04
     tantrums
    1.04
     angst
    0.93
     conspiracies
    0.91
     sarcasm
    0.90
     procrastination
    0.87
     hysteria
    0.85
     anecdotes
    0.85
     excuses
    0.84
    Act Density 1.184%

    No Known Activations