analiza aiqbench
This research introduces a novel framework for selecting Large Language Model teams by quantifying heterogeneity through error decorrelation and predictive divergence, moving beyond simple quality metrics to optimize system robustness. By formulating team composition as a combinatorial objective, the study demonstrates that strategically diverse model ensembles consistently outperform homogeneous baselines in reducing co-failures. These findings provide actionable principles for building more reliable multi-agent systems without relying on ad-hoc heuristics.
Which Models Work Well Together? Measuring Heterogeneity for LLM Team Selection
arXiv:2609.38274v1 Announce Type: new Abstract: The performance ceiling of an LLM team is constrained not only by individual model capabilities, but also by inter-member error resonance and predictive differences. Although heterogeneous teaming is often observed to be effective in practice, existing approaches lack complementarity metrics that are computable, interpretable, and optimizable, leaving team composition to rely on heuristics. We propose a heterogeneity-driven team selection framework that performs offline profiling to characterize individual capability along with two complementary signals: one captures decorrelation in error patterns to reduce co-failures, while the other measures divergence in predictive behavior to capture strategy diversity. We formulate team selection as a standardized quality--complementarity combinatorial objective and apply an efficient greedy search to select a small team from a candidate pool. Experiments across multiple benchmarks demonstrate that our framework consistently outperforms quality-only baselines under controlled candidate pools and team sizes, establishing reusable selection principles for multi-LLM systems.
Title:Which Models Work Well Together? Measuring Heterogeneity for LLM Team Selection
Abstract:The performance ceiling of an LLM team is constrained not only by individual model capabilities, but also by inter-member error resonance and predictive differences. Although heterogeneous teaming is often observed to be effective in practice, existing approaches lack complementarity metrics that are computable, interpretable, and optimizable, leaving team composition to rely on heuristics. We propose a heterogeneity-driven team selection framework that performs offline profiling to characterize individual capability along with two complementary signals: one captures decorrelation in error patterns to reduce co-failures, while the other measures divergence in predictive behavior to capture strategy diversity. We formulate team selection as a standardized quality--complementarity combinatorial objective and apply an efficient greedy search to select a small team from a candidate pool. Experiments across multiple benchmarks demonstrate that our framework consistently outperforms quality-only baselines under controlled candidate pools and team sizes, establishing reusable selection principles for multi-LLM systems.
Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA)
arXiv:2609.38274 [cs.CL]
(or arXiv:2609.38274v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2609.38274 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.
Text integral sub licența deschisă a publicației · 257 cuvinte
Subiecte conexe
Lessons from scaling teams of coding agents: reach the frontier at 30% of the cost
Cohere · 14 oct. 2026
How Sonos rebooted itself
The Verge · 3 sept. 2026
Biomedical Imaging's Real Bottleneck Is the Data, Not the Model
Databricks · 8 oct. 2026
Choosing Data Governance Tools for Enterprise Data Governance
Databricks · 25 aug. 2026