Tag
AOS-Bench, from Tsinghua University and Columbia researchers, evaluates LLM coding agents as workload managers handling streams of tasks under shared CPU and memory limits, finding that isolated task success doesn't carry over to streaming scenarios and that simple fixed-concurrency policies often match elaborate agent-managed ones.
Opti-Q is a database-inspired optimizer for multi-LLM question answering that plans execution DAGs to optimize answer quality under cost, latency, and energy constraints, achieving significant improvements on benchmarks.
This paper introduces BAGEN, a framework for evaluating budget awareness in LLM agents, defining budget estimation as internal and external budgets and formalizing progressive interval estimation. Experiments show that strong agents lack budget awareness, are over-optimistic, and that early stopping can save tokens while training improves alerting behavior.
Introduces TRIAGE, a framework for evaluating LLMs' prospective metacognitive control under token budgets, finding substantial gaps in their ability to allocate compute efficiently across problems.
This paper explores collaborative intelligence paradigms where distributed Large Language Models work together across devices and clouds to handle resource constraints. It covers vertical device-cloud collaboration, horizontal multi-agent collaboration, routing policies, and open research challenges in scalable and trustworthy cooperative AI.