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

    Neuronpedia

    INDEX
    Explanations

    socially unacceptable

    np_max-act-logits · gemini-2.5-flash-lite
    New Auto-Interp
    Negative Logits
     Notably
    0.57
    Notably
    0.56
     Certainly
    0.51
    確かに
    0.49
    Certainly
    0.48
     certainly
    0.47
     importantly
    0.46
     insgesamt
    0.45
     Importantly
    0.45
     Undoubtedly
    0.44
    POSITIVE LOGITS
     tantamount
    0.82
     indicative
    0.75
     heresy
    0.73
     acceptable
    0.71
     happening
    0.68
     unacceptable
    0.64
     taboo
    0.64
     unthinkable
    0.64
     disrespectful
    0.63
     done
    0.61
    Act Density 0.564%

    No Known Activations