ReguSim: Evaluating LLM Agent Rule Grounding in Financial Compliance
On the map: LLM agents and tool use → Financial trading and forecasting agents
Abstract (arXiv)
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.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Financial trading and forecasting agents, one of 16 directions in LLM agents and tool use. Builds on 17 of its 41 references that are in the map. First posted on arXiv 20 Aug 2026.
Builds on
17 of 41 references.
Closest ancestors · 6
Same direction, newest first
- 2026Behavioral Consistency Validation for LLM Agents: An Analysis of Trading-Style Switching through Stock-Market Simulation2 Feb
- 2026Replayable Financial Agents: A Determinism-Faithfulness Assurance Harness for Tool-Using LLM Agents17 Jan
- 2025StockSim: A Dual-Mode Order-Level Simulator for Evaluating Multi-Agent LLMs in Financial Markets12 Jul
- 2025Hide-and-Shill: A Reinforcement Learning Framework for Market Manipulation Detection in Symphony-a Decentralized Multi-Agent System12 Jul
- 2024TradingAgents: Multi-Agents LLM Financial Trading Framework28 Dec
- 2024Large Language Model Agent in Financial Trading: A Survey26 Jul
Other references · 11
- 2023BloombergGPT: A Large Language Model for Financecited by 923
- 2021LexGLUE: A Benchmark Dataset for Legal Language Understanding in Englishcited by 302
- 2021CUAD: An Expert-Annotated NLP Dataset for Legal Contract Reviewcited by 232
- 2021FinRL: Deep Reinforcement Learning Framework to Automate Trading in Quantitative Financecited by 57
- 2025TwinMarket: A Scalable Behavioral and Social Simulation for Financial Marketscited by 46
- 2026From Agent Traces to Trust: A Survey of Evidence Tracing and Execution Provenance in LLM Agentscited by 29
- 2024TradExpert: Revolutionizing Trading with Mixture of Expert LLMscited by 11
- 2025Agent-Based Simulation of a Financial Market with Large Language Modelscited by 6
- 2023Detecting Pump&Dump Stock Market Manipulation from Online Forumscited by 3
- 2021Protecting Retail Investors from Order Book Spoofing using a GRU-based Detection Modelcited by 2
- 2026Legal-DC: Benchmarking Retrieval-Augmented Generation for Legal Documentscited by 2
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Measurement Boundaries in LLM Financial Agent Evaluation: Fixed-Tape Execution and Multi-Defect Auditing
- 2026Beyond Agent Architecture: Execution Assumptions and Reproducibility in LLM-Based Trading Systems
- 2025Standard Benchmarks Fail -- Auditing LLM Agents in Finance Must Prioritize Risk
- 2026FinHarness: An Inline Lifecycle Safety Harness for Finance LLM Agents
- 2026CLQT: A Closed-Loop, Cost-Aware, Strategy-Consistent Benchmark for Diagnostic Evaluation of LLM Portfolio-Management Agents
- 2026Can Blindfolded LLMs Still Trade? An Anonymization-First Framework for Portfolio Optimization
- 2026The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applications
- 2026Agentic Trading: When LLM Agents Meet Financial Markets