MidTool: Mid-training Data Synthesis for Agentic Tool Use
Mid-training is increasingly recognized as a critical stage for shaping the capabilities of large language models. Recent work has shown that targeted mid-training can strengthen reasoning-intensive abilities such as math and science, and can also improve agentic capabilities in software-engineering settings. In this work, we study the parallel but less explored agentic capability: general tool use. We present MidTool, an open corpus construction pipeline for agentic tool-use mid-training that combines large-scale web, PDF, and code data with synthesized supervision from real-world tool APIs,
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- LinkedLinked via arxiv author · 85%Fengqing Jiang →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Yite Wang →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Boyi Liu →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Zhaoyang Wang →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Canwen Xu →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Zhewei Yao →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Radha Poovendran →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”
- LinkedLinked via arxiv author · 85%Yuxiong He →
“MidTool: Mid-training Data Synthesis for Agentic Tool Use”