Metraipolis › archive › dc_80058053233d

CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling — arXiv:2610.06829

Source
arXiv
Original
https://arxiv.org/abs/2610.06829
License (as archived)
CC0-1.0
Archived
2026-10-06T05:27:54.437Z
SHA-256 of the archived text
8decacac678de67d37a0c57c45d701f687a6869925dbd23be9e9d63b3367187e
Stable URL of this page
https://metraipolis.com/documento/dc_80058053233d

How to cite

CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling — arXiv:2610.06829 — arXiv, https://arxiv.org/abs/2610.06829 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_80058053233d (Metraipolis archive dc_80058053233d, sha256 8decacac678de67d37a0c57c45d701f687a6869925dbd23be9e9d63b3367187e)
BibTeX
@misc{dc80058053233d,
  title = {CLIFT: Conformal Self-Verification for Web Agent Training and Test-Time Scaling — arXiv:2610.06829},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06829},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_80058053233d; sha256 8decacac678de67d37a0c57c45d701f687a6869925dbd23be9e9d63b3367187e},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06829. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Open-source web agents are now strong enough to execute realistic browser tasks, but training them with reinforcement learning still depends on weak supervision: binary task success is too sparse for credit assignment, while frontier-language-model judges are too expensive to call at every step and cannot be assumed available at deployment. We introduce CLIFT, a training and test-time scaling method built around conformal self-verification. During training, the agent answers natural-language verification questions about its own rollouts; a Compositional Conformal Certifier keeps only question signals whose URL-conditional evidence agrees with a training-time judge, assigns signed trust weights through polarity-aware lift, and blends the resulting verifier score into per-step rewards in a way that never subtracts from the judge baseline. At test time, the same certified bank is frozen and reused as structured evidence for Conformal Trajectory Selection (CTS): the agent samples a greedy rollout and one or more diverse retries, the self-verifier summarises each URL trace, and a conservative majority-vote rule chooses whether to swap away from the current incumbent without calling any external judge. This single mechanism supports three settings. On WebArena Infinity, CLIFT achieves state-of-the-art performance among open-source web agents. On VisualWebArena, a bank trained with the open model transfers to GPT-5.5 at test time and reaches state-of-the-art performance under the canonical harness. On Online Mind2Web, without training an agent on the benchmark, translating the certified question bank improves a live-web agent in zero-shot evaluation. Together these results position conformal self-verification as a way to turn costly judge feedback into a reusable training signal and a judge-free test-time scaling signal.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.06829 — 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 2026-10-06T05:24:06.197Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).