Preprint
Jul 2026
AMTFV: Agentic Mathematical Tool-Flow Verification for LLM Self-Correction
Experimental results show that AMTFV outperforms the representative baselines evaluated in this study overall; under an individual model configuration, it improves average accuracy over the strongest baseline by up to 8.3 percentage points.
Rui-Zheng Zou, Yutao Zhu, Mengqi Wei et al.
· 0 citations