Small language models (SLMs) are attractive backbones for agentic systems, but their limited parametric capacity increases their reliance on reliable tool use and, consequently, on training data containing tool interactions that actually execute. Query-first synthesis can produce plausible requests that do not correspond to valid tool sequences, compatible parameters, or available data. We propose SYNTHETICAGENTTRACEQA, an execution-first framework that constructs workflow templates, assigns tools under data-flow constraints, executes and validates traces, and only then synthesizes user tasks, teacher reasoning, and reference answers. Across four tool ecosystems, fine-tuning Qwen3.5-4B and 9B on the resulting data improves tool execution, reference-trace agreement, and answer production (numeric match remains <10%). In our experiments, the best results came from training only on trajectories that executed successfully and supervising tool calls and final answers while masking teacher
Execution-first synthetic tool-use traces for small language model agents
SLM-Agents, 1st Workshop on SLMs for Agentic Systems, in 40th Conference on Neural Information Processing Systems (NeurIPS 2026), 13 December 2026, Paris, France
Type:
Conference
City:
Paris
Date:
2026-12-13
Department:
Data Science
Eurecom Ref:
9056
See also:
PERMALINK : https://www.eurecom.fr/publication/9056