This is not a yet known terminology , but I suggest it for Poulet-numbers with the property that they give another Poulet-number , if the decimal expansion is written down in reverse order analogue to the emirp's (primes that given another prime if their decimal expansion is written down in reverse).
A poulet-number is a composite number $N$ satisfying the congruence $2^{N-1}\equiv 1\mod N$ , so it passes the weak $2$-fermat-test.
A "Teluop-number" is called "trivial" if it is a palindrome (the deicmal expansion written in reverse order gives the same number)
There are $5$ examples below $10^5$ :
gp > forstep(n=3,10^8,2,if(Mod(2,n)^(n-1)==1,if(ispseudoprime(n)==0,m=fromdigits(Vecrev(digits(n)));if(Mod(2,m)^(m-1)==1,if(ispseudoprime(m)==0,print(n))))))
15709
90751
101101
129921
1837381
gp >
$$127665878878566721$$ is a palindrome Carmichael-number.
In the database of the Carmichael-numbers upto $2^{64}$ , there is no other such entry , only $101101$ and the large number given above. There is no "nontrivial" example. In total , we have four trivial "Teluop-numbers" and one nontrivial pair (the reverse of a Teluop-number is of course a Teluop-number).
I have no access to the database of the Poulet-numbers upto $2^{64}$.
Are there more examples in the database ? Is the number of Teluop-numbers infinite ? I think not because Poulet-numbers get very rare if the magnitude gets over , say , $10^{20}$.
In particular interesting would be a nontrivial Carmichael-number with the above property.