CauSim: Scaling Causal Reasoning with Increasingly Complex Causal Simulators

ArXi:2605.09079v1 Announce Type: new Despite surpassing human performance across mathematics, coding, and other knowledge-intensive tasks, large language models (LLMs) continue to struggle with causal reasoning. A core obstacle is the target data itself: causal systems are complex and often expressed in non-executable forms, while ground-truth answers to causal queries are inherently scarce. We