SOURCE-LINKED INTELLIGENCE

Monitorable Chart Reasoning Agents via Verifiable Process Rewards

arXiv · AI, language, vision and robotics · article · Sep 21, 2026 · UTC

Chart reasoning agents are increasingly used to extract actionable insights in critical domains, achieving state-of-the-art performance on multiple benchmarks. Yet, high benchmark accuracy alone is insufficient for deployment, where stakeholders must be able to audit and verify how a model reaches its answer. Existing LVLM-based chart agents produce either answer-only predictions or free-form rationales that are hard to verify, obscuring whether an error arose from misreading the chart, extracting a wrong value, or miscomputing. We propose Chart-RVR, a reinforcement learning framework for trai

Read original source ↗ Open in workspace

recordType

paper

region

Global