{
  "annotation_id": "fe049dcd-02d9-4a80-b00b-352db94f4d53",
  "date_created": "2026-02-17T05:53:16.276000Z",
  "date_modified": "2026-02-17T05:53:16.276000Z",
  "file_hash": "65c29dfb24985ab541ee0e7d0b02460b58d9f28410e510683d05ae5a7e774159",
  "private": false,
  "record": {
    "abstract": "Language models now provide an interface to express and often solve general problems in natural language, yet their ultimate computational capabilities remain a major topic of scientific debate. Unlike a formal computer, a language model is trained to autoregressively predict successive elements in human-generated text. We prove that chaining a language model\u0027s autoregressive output is sufficient to perform universal computation. That is, a language model can simulate the execution of any algorithm on any input. The challenge of eliciting desired computational behaviour can thus be reframed in terms of programmability: the ease of finding a suitable prompt. Strikingly, we demonstrate that even randomly initialized language models are capable of universal computation before training. This implies that training does not give rise to computational expressiveness -- rather, it improves programmability, enabling a natural language interface for accessing these intrinsic capabilities.",
    "arxiv_id": "2601.08061",
    "authors": [
      "Alex Lewandowski",
      "Marlos C. Machado",
      "Dale Schuurmans"
    ],
    "categories": [
      "cs.CL"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "title": "Universal computation is intrinsic to language model decoding",
    "url": "https://arxiv.org/abs/2601.08061",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "5def9480-7883-4a1c-ae68-81ece32457e2",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}