Metraipolis › archive › dc_29eaf48015bb
Efficient Scalable Oversight via Rewinding — arXiv:2610.40535
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.40535
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:10:51.074Z
- SHA-256 of the archived text
9bc8a8995679e2acc0eadf5c0da64b64f94c2af2d120a3154dc39afc9e8ffec4- Stable URL of this page
https://metraipolis.com/documento/dc_29eaf48015bb
How to cite
Efficient Scalable Oversight via Rewinding — arXiv:2610.40535 — arXiv (en), https://arxiv.org/abs/2610.40535 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_29eaf48015bb (Metraipolis archive dc_29eaf48015bb, sha256 9bc8a8995679e2acc0eadf5c0da64b64f94c2af2d120a3154dc39afc9e8ffec4)
BibTeX
@misc{dc29eaf48015bb,
title = {Efficient Scalable Oversight via Rewinding — arXiv:2610.40535},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.40535},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_29eaf48015bb; sha256 9bc8a8995679e2acc0eadf5c0da64b64f94c2af2d120a3154dc39afc9e8ffec4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.40535. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Scalable oversight asks how a weak verifier can check the outputs of computations performed by more powerful AI systems. One proposed approach uses relativizing interactive proofs, in which a weak verifier checks the output of a complex oracle-aided computation by interacting with one or more powerful provers. A prominent example is debate, which is highly efficient but assumes that one of two powerful AI models, or provers, is truthful. For certain classes of oracle-aided computations, this assumption can be removed, but at substantially greater computational cost. In this work, we introduce \emph{rewindable interactive proofs} (RIPs), a new model of relativizing interactive proofs for scalable oversight. In an RIP, the verifier may rewind the prover to an earlier state of the interaction and continue from there. This is possible when the prover is an AI model, since the verifier can save a snapshot of the model's state and later restore it. We construct a relativizing RIP for computations whose outputs are robust to changing a small fraction of oracle answers. Our construction is more efficient than prior relativizing proofs that do not assume a truthful prover, and it uses no cryptography. In our construction the honest prover runs in time linear in that of the original computation, and the verifier, after a one-time preprocessing step, runs in time nearly linear in the input length (which can be much smaller than the computation itself). We implement our protocol, report a preliminary MNIST experiment, and show that an RIP can be faster than a state-of-the-art deployed proof system. RIPs may be of independent interest. Rewinding is possible whenever the verifier controls the prover's machine, for example when the prover runs on trusted hardware or in a virtual machine whose state can be checkpointed.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computational Complexity (cs.CC); Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2610.40535 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T06:07:08.567Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).