Looped Transformers as Programmable Computers · HackerPapers discussions

Looped Transformers as Programmable Computers — HN submission

Thu, 10 Sep 2026 02:46:31 GMT

Looped Transformers as Programmable Computers (2023)

4 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.

Looped Transformers as Programmable Computers — mentioned by stavros

Wed, 13 Dec 2023 07:48:06 GMT

Yep: https://arxiv.org/pdf/2301.13196.pdf

Paper and discussion history. A mention is not an endorsement.

Looped Transformers as Programmable Computers — mentioned by nojvek

Thu, 15 Jun 2023 18:51:36 GMT

As others have mentioned, turing machines can be achieved by only one instruction set. See https://en.wikipedia.org/wiki/One-instruction_set_computer Most popular being subleq (subract if less than or equal to zero). Subleq can be implemented with a transformer. See https://arxiv.org/pdf/2301.13196.pdf (Looped Transformers as Programmable Computers). The important distinction is that transformers need to be looped. Just like ChatGPT calls the model over and over again until stop token or limit is reached. And we know ChatGPT does better if we let it think step by step and blabber more (chain…

Paper and discussion history. A mention is not an endorsement.