OrchDAG: Complex Tool Orchestration in Multi-Turn Interactions with Plan DAGs

Yifu Lu, Shengjie Liu, Li Dong

Keywords: Multi-turn, Planning, RLVR

TL;DR: OrchDAG introduces a synthetic DAG-based data generation pipeline and graph-reward framework that benchmarks and improves multi-turn tool use in RLVR, offering a challenging yet solvable testcase for agentic tool interactions.

Abstract:

Agentic tool use has gained traction with the rise of agentic tool calling, yet
most existing work overlooks the complexity of multi-turn tool interactions. We
introduce OrchDAG, a synthetic data generation pipeline that models tool execution
as directed acyclic graphs (DAGs) with controllable complexity. Using this dataset,
we benchmark model performance and propose a graph-based reward to enhance
RLVR training. Experiments show that the dataset presents a challenging but
solvable benchmark, and the proposed reward is effective when combined with
GRPO-style algorithms, highlighting the importance of leveraging topological
structure and data complexity in multi-turn tool use.

Submission Number: 118