> almost 100% accuracy I think it's annoying that this gets repeated without revealing the actual stats: this is for the BFI test, it's 99.5% ± 0.38, where 40

> almost 100% accuracy I think it's annoying that this gets repeated without revealing the actual stats: this is for the BFI test, it's 99.5% ± 0.38, where 40 samples per person were split 80-20 in training and testset. paper: https://dl.acm.org/doi/abs/10.1145/3719027.3765062 https://dl.acm.org/doi/abs/10.1145/3719027.3765062