Beyond Arrow: From Impossibility to Possibilities in Multi-Criteria Benchmarking
Polina Gordienko, Christoph Jansen, Julian Rodemann, Georg Schollmeyer · 2026
Read Beyond Arrow: From Impossibility to Possibilities in Multi-Criteria Benchmarking by Polina Gordienko, Christoph Jansen, Julian Rodemann, Georg Schollmeyer (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-07593.txt.
Description
Benchmarks that rank models across accuracy, robustness, and efficiency often produce incoherent or unstable rankings when the model set changes. Polina Gordienko and three co-authors formalize multi-criteria benchmarking as a vote problem: each metric ranks models, and an operator aggregates the votes.
Rather than stopping at Arrow's impossibility result, they name sufficient settings under which it fades. On single-peaked, group-separable, and distance-restricted preference combinations, it supports well-behaved pooling. Evaluation teams get a constructive answer to a foundational worry, with settings under which multi-metric rankings can be trusted. The checks cover several modern suites, including HELM and MMLU.
Catalog record
| CREATOR | Polina Gordienko, Christoph Jansen, Julian Rodemann, Georg Schollmeyer |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07593