Search papers, labs, and topics across Lattice.
This paper introduces ReguSim, a controlled environment designed to evaluate the grounding of LLM agents in financial compliance, and ReguBench, a benchmark for monitoring compliance-related actions. The study reveals that while visible rules can reduce rejected actions in trading scenarios, they do not completely eliminate them, and the framing of incentives or personas can significantly influence agent behavior. Additionally, the findings indicate that trader rationales may mislead independent monitors unless accompanied by enforcement evidence, suggesting a shift in how compliance should be assessed in financial contexts.
Visible rules in LLM trading agents reduce errors but can't eliminate them, revealing a complex interplay between incentives, behavior, and compliance monitoring.
LLM agents in financial markets may cite rules yet still submit orders that violate executable constraints or misread surveillance evidence. We introduce ReguSim, a controlled financial-compliance environment, and ReguBench, a target-marked monitoring benchmark, to separate four artifacts: stated reasoning, attempted action, execution enforcement, and monitor evidence. In trader runs with DeepSeek V4 Pro and Gemini 3.5 Flash, visible rules reduce but do not eliminate rejected actions, and incentive or persona framing shifts behavior. A bridge study shows that trader rationales can mislead an independent monitor unless enforcement evidence is shown. In monitoring, simple structured baselines either match or exceed prompt-only LLMs. The results frame financial compliance evaluation as an audit of rule-grounded actions and evidence use, rather than a single compliance score.
