| title | Looped Language Models Improve Compositional Tool Calling |
|---|---|
| authors | Andrei Cristian Popescu, Haitz Sáez de Ocáriz Borde, Pietro Liò |
| date | 2026-08-17 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2608.18171 |
| upvotes | 14 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf |
| abstract | Looped language models have shown promising results on reasoning benchmarks, yet their potential for agentic tool use remains largely unexplored. We study this question in compositional tool-calling settings, where models must coordinate multiple API calls, maintain intermediate state, and preserve dependencies across tool interactions. We evaluate native and retrofitted looped language models on API-Bank, BFCL, and NESTful, comparing looped and non-looped models trained under matched supervised fine-tuning recipes and varying recurrent depth at inference time. In controlled experiments, recurrent computation generally benefits compositional and dependency-aware tool use, while providing smaller and more model-dependent gains on isolated API invocation. Accuracy on multi-step tool use generally increases with recurrent depth; adaptive inference, however, achieves a more favorable compute-performance trade-off by allocating additional computation only when needed. Our results suggest that looped language models are a promising architecture for agentic systems that require reliable planning, coordination, and execution of compositional tool use workflows. |
cite
@misc{popescu2026looped,
title={Looped Language Models Improve Compositional Tool Calling},
author={Andrei Cristian Popescu and Haitz Sáez de Ocáriz Borde and Pietro Liò},
year={2026},
eprint={2608.18171},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2608.18171}
}