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

    Neuronpedia

    INDEX
    Explanations

    phrases and sentences related to contrast or contradiction

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of phrases that involve conditional or hypothetical scenarios

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Negative Logits
     PLUS
    -0.90
    anwhile
    -0.76
    fleet
    -0.75
    .?
    -0.70
     srfAttach
    -0.68
     GOODMAN
    -0.66
    ļéĨĴ
    -0.66
    EStream
    -0.65
     sqor
    -0.63
    INS
    -0.63
    POSITIVE LOGITS
     technically
    1.24
     admittedly
    1.16
     laud
    1.12
     undoubtedly
    1.09
     ostensibly
    1.08
     initially
    1.05
     certainly
    1.00
     theoretically
    0.98
     admirable
    0.96
     undeniably
    0.94
    Act Density 0.378%

    No Known Activations