why compute bernoulli entropy in this way?

[lucas-yang256]

as the code written [here](def logit_bernoulli_entropy(logits_B):
ent_B = (1.-tensor.nnet.sigmoid(logits_B))*logits_B - logsigmoid(logits_B)
return ent_B), bernoulli was computed by this way

imitation/policyopt/thutil.py

Lines 48 to 51 in 99fbccf

	def logit_bernoulli_entropy(logits_B):
	ent_B = (1.-tensor.nnet.sigmoid(logits_B))*logits_B - logsigmoid(logits_B)
	return ent_B

but it's different to the equation of binary entropy:
$-p\log p - (1-p)\log(1-p)$

is there any relationship between these two expressions? or why does openai compute bernoulli entropy that way? is there any theoretical equation support?