top | item 47210374 (no title) thesz | 5 hours ago Transformers are able to recognize balanced brackets grammar at 97% success rate: https://openreview.net/pdf?id=kaILSVAspnThis is 3% or infinitely far away from the perfect tech.The perfect tech is the stack. discuss order hn newest krackers|1 hour ago This is very interesting since there is another notable paper which shows LLMs can recognize and generate CFGshttps://arxiv.org/abs/2305.13673and of course a^n b^n is also classic CFG, so it's not clear why one paper had positive results while the other hand negative. thesz|1 hour ago Dyck grammar (balanced brackets) are not an a^nb^n, there are several kinds of brackets.I cannot find probability of success in paper you linked. Is it 100%? I believe it is less than 100%, because LLMs are intrinsically probabilistic machines. load replies (1)
krackers|1 hour ago This is very interesting since there is another notable paper which shows LLMs can recognize and generate CFGshttps://arxiv.org/abs/2305.13673and of course a^n b^n is also classic CFG, so it's not clear why one paper had positive results while the other hand negative. thesz|1 hour ago Dyck grammar (balanced brackets) are not an a^nb^n, there are several kinds of brackets.I cannot find probability of success in paper you linked. Is it 100%? I believe it is less than 100%, because LLMs are intrinsically probabilistic machines. load replies (1)
thesz|1 hour ago Dyck grammar (balanced brackets) are not an a^nb^n, there are several kinds of brackets.I cannot find probability of success in paper you linked. Is it 100%? I believe it is less than 100%, because LLMs are intrinsically probabilistic machines. load replies (1)
krackers|1 hour ago
https://arxiv.org/abs/2305.13673
and of course a^n b^n is also classic CFG, so it's not clear why one paper had positive results while the other hand negative.
thesz|1 hour ago
I cannot find probability of success in paper you linked. Is it 100%? I believe it is less than 100%, because LLMs are intrinsically probabilistic machines.