Skip to content
aiqbench takeAI Safety

aiqbench analysis

This research introduces OfficeEditBench, a rigorous benchmark evaluating AI models' ability to perform constrained edits on Office files while preserving dependent logic and protected states. The study reveals that current systems fail to satisfy complete task contracts, often breaking formulas or omitting prerequisites despite achieving superficial delivery rates. This work advances the field of AI safety and reliability by defining new standards for artifact-level maintenance in complex document ecosystems.
Source coveringarXiv
What this isA model's reading of the article and related coverage, written for aiqbench. Not the publisher's words.

The Last Mile Is the File: OfficeEditBench for Preservation-Aware Office Editing

arXiv2 min read

arXiv:2609.34469v1 Announce Type: cross Abstract: A small Office edit creates two obligations: propagate every required update and leave protected state untouched. Updating too little leaves dependencies inconsistent; updating too much changes content the user did not authorize. We introduce OfficeEditBench, a 170-task benchmark for change-scoped maintenance of spreadsheets, presentations, and documents. Task contracts specify required updates, protected state, native structures, and applicable interaction requirements. Across 510 archived task-system outcomes from WorkBuddy, Doubao, and Codex, we distinguish file delivery, target completion, and verifier-defined acceptance. Hard package-valid delivery ranges from 92% to 100%, yet no selected output satisfies the complete contract. Case analysis highlights why local correctness is insufficient: an updated value can lose its generating formula, a revised rule can fail to reach related conclusions, and a new deadline can omit a retained prerequisite. These mechanisms connect artifact-level checks to the continued maintainability of Office files. We analyze maintenance failures while distinguishing frozen automatic verdicts from human acceptability. OfficeEditBench provides a testbed for completing required changes while preserving the logic and scope of existing work.

Abstract:A small Office edit creates two obligations: propagate every required update and leave protected state untouched. Updating too little leaves dependencies inconsistent; updating too much changes content the user did not authorize. We introduce OfficeEditBench, a 170-task benchmark for change-scoped maintenance of spreadsheets, presentations, and documents. Task contracts specify required updates, protected state, native structures, and applicable interaction requirements. Across 510 archived task-system outcomes from WorkBuddy, Doubao, and Codex, we distinguish file delivery, target completion, and verifier-defined acceptance. Hard package-valid delivery ranges from 92% to 100%, yet no selected output satisfies the complete contract. Case analysis highlights why local correctness is insufficient: an updated value can lose its generating formula, a revised rule can fail to reach related conclusions, and a new deadline can omit a retained prerequisite. These mechanisms connect artifact-level checks to the continued maintainability of Office files. We analyze maintenance failures while distinguishing frozen automatic verdicts from human acceptability. OfficeEditBench provides a testbed for completing required changes while preserving the logic and scope of existing work.

23 pages, 7 figures. Benchmark and code: this https URL

Software Engineering (cs.SE); Computation and Language (cs.CL)

arXiv:2609.34469 [cs.SE]

(or arXiv:2609.34469v1 [cs.SE] for this version)

https://doi.org/10.48550/arXiv.2609.34469 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Full text under the publisher's open licence · 277 words

Related coverage