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

    Neuronpedia

    INDEX
    Explanations

    He said, thinks, chuckles

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
    Statement
    0.45
     Presumably
    0.45
     ostensibly
    0.41
     Statement
    0.40
     Establish
    0.39
     Statements
    0.39
     egreg
    0.39
     excessively
    0.39
     установ
    0.38
    おそらく
    0.38
    POSITIVE LOGITS
     admits
    0.86
     thinks
    0.84
     chuckled
    0.84
     chuckle
    0.79
     admitted
    0.77
     concedes
    0.77
     said
    0.73
     laughed
    0.72
     laugh
    0.70
     joked
    0.70
    Act Density 0.004%

    No Known Activations