Metraipolis › archive › dc_b11f557db436
Beyond Scalar Leaderboards: Adaptive Sampling for Pareto Frontier Identification in Multi-Objective LLM Arenas — arXiv:2610.27447
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27447
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:07:45.135Z
- SHA-256 of the archived text
eaa1a788ff9a192ebc4291eda1c3672846a9b00ef6af2bd81d55ebbca825030b- Stable URL of this page
https://metraipolis.com/documento/dc_b11f557db436
How to cite
Beyond Scalar Leaderboards: Adaptive Sampling for Pareto Frontier Identification in Multi-Objective LLM Arenas — arXiv:2610.27447 — arXiv (en), https://arxiv.org/abs/2610.27447 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b11f557db436 (Metraipolis archive dc_b11f557db436, sha256 eaa1a788ff9a192ebc4291eda1c3672846a9b00ef6af2bd81d55ebbca825030b)
BibTeX
@misc{dcb11f557db436,
title = {Beyond Scalar Leaderboards: Adaptive Sampling for Pareto Frontier Identification in Multi-Objective LLM Arenas — arXiv:2610.27447},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27447},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b11f557db436; sha256 eaa1a788ff9a192ebc4291eda1c3672846a9b00ef6af2bd81d55ebbca825030b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27447. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
The evaluation of large language models (LLMs) is inherently multi-objective, as different models may offer different tradeoffs across performance dimensions such as math, code, and writing. Because such evaluations are often based on costly and time-consuming human feedback, it is important to draw reliable conclusions using as few comparisons as possible. In this paper, we present Vector-Borda Empirical-Gap Elimination (VB-EGE), a fixed-confidence adaptive sampling algorithm for efficiently identifying models on the Pareto frontier in arena-style evaluation, where each comparison evaluates a pair of models on a prompt with respect to a selected performance dimension. VB-EGE progressively eliminates models from an active set once sufficient statistical evidence has been accumulated to certify their Pareto status. Under a multi-objective Bradley-Terry model, we show that VB-EGE returns the correct Pareto frontier with high probability and derive an instance-dependent upper bound on its sample complexity. We further establish a worst-case lower bound showing that VB-EGE is minimax-optimal up to logarithmic factors. Simulations and experiments on three real-world Arena datasets (Arena, SearchArena, and VisionArena) demonstrate VB-EGE's empirical effectiveness. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Machine Learning (stat.ML) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Applications (stat.AP); Other Statistics (stat.OT). Pagina del lavoro: https://arxiv.org/abs/2610.27447 — 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 2026-10-21T07:05:43.560Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).