Search papers, labs, and topics across Lattice.
MilaCalifornia State UniversityCharlieColumbiaGeorgia TechMBZUAIMcGillThe Fin AIFeb 19, 2026arXiv:2602.16990The paper introduces Conv-FinRe, a new benchmark for financial stock recommendation that evaluates LLMs based on both behavioral alignment and utility-grounded decision quality, addressing the limitations of solely imitating user behavior which can be noisy and short-sighted. Conv-FinRe provides multi-view references that distinguish descriptive behavior from normative utility based on investor-specific risk preferences, allowing for diagnosis of rational analysis, noise imitation, or market momentum influence. Experiments with state-of-the-art LLMs reveal a trade-off between rational decision quality and behavioral alignment, highlighting the challenge of balancing utility-based ranking with matching user choices.
LLMs struggle to balance rational financial decisions with mimicking noisy user behavior, often overfitting to short-term market trends instead of aligning with long-term investment goals.
Most recommendation benchmarks evaluate how well a model imitates user behavior. In financial advisory, however, observed actions can be noisy or short-sighted under market volatility and may conflict with a user's long-term goals. Treating what users chose as the sole ground truth, therefore, conflates behavioral imitation with decision quality. We introduce Conv-FinRe, a conversational and longitudinal benchmark for stock recommendation that evaluates LLMs beyond behavior matching. Given an onboarding interview, step-wise market context, and advisory dialogues, models must generate rankings over a fixed investment horizon. Crucially, Conv-FinRe provides multi-view references that distinguish descriptive behavior from normative utility grounded in investor-specific risk preferences, enabling diagnosis of whether an LLM follows rational analysis, mimics user noise, or is driven by market momentum. We build the benchmark from real market data and human decision trajectories, instantiate controlled advisory conversations, and evaluate a suite of state-of-the-art LLMs. Results reveal a persistent tension between rational decision quality and behavioral alignment: models that perform well on utility-based ranking often fail to match user choices, whereas behaviorally aligned models can overfit short-term noise. The dataset is publicly released on Hugging Face, and the codebase is available on GitHub.
