# COVER: Identifiable Evaluation of Coalition Routing

arXiv: https://arxiv.org/abs/2608.28475 · HTML: https://mlsanity.com/paper/2608.28475/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 20 references that are in the map.
- First posted on arXiv 28 Aug 2026.

## Builds on (most cited first, 10 of 10)

- [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://mlsanity.com/paper/2308.08155/) (2023)
- [ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs](https://mlsanity.com/paper/2307.16789/) (2023)
- [Improving Factuality and Reasoning in Language Models through Multiagent Debate](https://mlsanity.com/paper/2305.14325/) (2023)
- [Deep Sets](https://mlsanity.com/paper/1703.06114/) (2017)
- [RouteLLM: Learning to Route LLMs with Preference Data](https://mlsanity.com/paper/2406.18665/) (2024)
- [Mixture-of-Agents Enhances Large Language Model Capabilities](https://mlsanity.com/paper/2406.04692/) (2024)
- [A Dynamic LLM-Powered Agent Network for Task-Oriented Agent Collaboration](https://mlsanity.com/paper/2310.02170/) (2023)
- [Effective Off-Policy Evaluation and Learning in Contextual Combinatorial Bandits](https://mlsanity.com/paper/2408.11202/) (2024)
- [Leveraging Factored Action Spaces for Off-Policy Evaluation](https://mlsanity.com/paper/2307.07014/) (2023)
- [Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation](https://mlsanity.com/paper/2608.20316/) (2026)

## Abstract (arXiv)

When a multi-agent system changes its team, it also changes the messages and final answer it produces, so an end-to-end accuracy gap does not by itself identify a routing effect. We introduce method, an evaluation contract that fixes a public information boundary, downstream stack G, and finite legal team family before outcomes are generated. Complete coverage identifies exact finite-benchmark oracle regret conditional on that stack. For any finite collection of frozen policies, executing the union of their distinct selected teams is the minimal assumption-free support for every pairwise policy contrast, though not for absolute oracle regret. Two controlled tables with source-ID-disjoint splits test the instrument. On MuSiQue-12, a pre-specified privileged positive control improves regret from 0.532 to 0.402; a later public-interface control reaches 0.424 versus 0.554 but is retrospective. On HotpotQA-4, a pre-specified public direct scorer improves regret from 0.313 to 0.110. In fixed-stack Llama execution, verified route regret improves by 0.190, while the raw-answer gain is 0.010 with an interval crossing zero. A five-family ToolSandbox variant-shift validation exhaustively evaluates 16 declared teams on 14 untouched task variants (224/224 valid rows): the declared-family oracle reaches 0.768 safe-evidence completion, while the prospectively frozen router gets 0.637 (regret 0.131), failing the predeclared 0.10 criterion. A later retrospective comparator reaches 0.655, matching all-workers with 4.57 versus 5.00 workers on average. Thus COVER exposes selection headroom without manufacturing a routing win. A crossed-stack diagnostic shows absolute scores depend on G but finds no detectable router-by-finalizer interaction. COVER is an auditable measurement methodology, not a claim of stack-invariant or universal agent-routing superiority.