Metraipolis › archive › dc_ad232bd3f40a

Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision — arXiv:2609.02057

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.02057
License (as archived)
CC0-1.0
Archived
2026-10-17T03:25:47.971Z
SHA-256 of the archived text
4ec7c2b8c3475aa4fb1fb0a6af673aa4f5a7966a926059f0737db73be0505e96
Stable URL of this page
https://metraipolis.com/documento/dc_ad232bd3f40a

How to cite

Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision — arXiv:2609.02057 — arXiv (en), https://arxiv.org/abs/2609.02057 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ad232bd3f40a (Metraipolis archive dc_ad232bd3f40a, sha256 4ec7c2b8c3475aa4fb1fb0a6af673aa4f5a7966a926059f0737db73be0505e96)
BibTeX
@misc{dcad232bd3f40a,
  title = {Monitoring Web Agents Without Internal Signals: Observable Trajectories and Key-Step Supervision — arXiv:2609.02057},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.02057},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ad232bd3f40a; sha256 4ec7c2b8c3475aa4fb1fb0a6af673aa4f5a7966a926059f0737db73be0505e96},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.02057. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Reliable web-agent monitoring is difficult when model-internal uncertainty signals such as token logits are unavailable. In this work, we study prefix-level risk prediction for web agents using observable trajectory signals: given an evolving prefix, estimate whether the current execution remains on track or is tending toward failure. We derive two observable trajectory representations: Macro features summarize cross-step agent--environment behavior and feedback, while Micro features measure the consistency of intention, action, and anticipated state change through repeated black-box queries. Instead of inheriting the final result label, we label the first critical error that remains uncorrected in the observed continuation and is associated with final failure as a key-step boundary, preserving valid early prefixes of failed trajectories as on track. Across WebArena-Lite and Online Mind2Web web agent benchmarks with five open- and closed-source backbones, observable trajectory signals are competitive with internal-signal baselines. The resulting predictors also support early intervention under fixed false-cut budgets and transfer across held-out website categories. These findings show that observable trajectory signals support valuable risk prediction abilities.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-03 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.02057 — 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 2026-10-17T03:13:12.413Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).