Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2510.05702

Uncovering Representation Bias for Investment Decisions in Open-Source Large Language Models

Fabrizio Dimino·Krati Saxena·Bhaskarjit Sarmah·Stefano Pasquali
View original source ↗Read source document ↗Export metadata ↓

Abstract

Large Language Models are increasingly adopted in financial applications to support investment workflows. However, prior studies have seldom examined how these models reflect biases related to firm size, sector, or financial characteristics, which can significantly impact decision-making. This paper addresses this gap by focusing on representation bias in open-source Qwen models. We propose a balanced round-robin prompting method over approximately 150 U.S. equities, applying constrained decoding and token-logit aggregation to derive firm-level confidence scores across financial contexts. Using statistical tests and variance analysis, we find that firm size and valuation consistently increase model confidence, while risk factors tend to decrease it. Confidence varies significantly across sectors, with the Technology sector showing the greatest variability. When models are prompted for specific financial categories, their confidence rankings best align with fundamental data, moderately with technical signals, and least with growth indicators. These results highlight representation bias in Qwen models and motivate sector-aware calibration and category-conditioned evaluation protocols for safe and fair financial LLM deployment.

Explore related subjects

q-fin.CPcs.AI

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Fabrizio Dimino, Krati Saxena, Bhaskarjit Sarmah, Stefano Pasquali. 2025-11-03. Uncovering Representation Bias for Investment Decisions in Open-Source Large Language Models. https://arxiv.org/abs/2510.05702

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Are Whitepaper Claims Reflected in Market Structure? A Contamination-Aware Pipeline and a Power-Limited Null

Do the functional narratives in cryptocurrency whitepapers correspond to their tokens' market behaviour? We compare ten-category topical-emphasis profiles for 43 screened documents with seven market statistics calculated from 2023--2024 exchange data. The primary specification reconstructs USD notional turnover from hourly bars. Dimension-matched Procrustes congruence is $ϕ=0.280$ (permutation $p=0.507$), slightly below its permutation-null mean of $0.284$; the zero-padded statistic gives the same non-detection. A four-leg comparison separates document replacement from changes in the assets included. Replacing documents on the 34 common assets changes padded congruence by $-0.014$ under USD turnover and $-0.009$ under base-token volume. Entity rankings and threshold crossings depend on both composition and specification, so an earlier contamination-only attribution is withdrawn. The documents are not a verified historical corpus: at least two postdate the market window. Excluding these documents, or excluding all seven assets with shorter histories, does not produce a significant alignment. Fresh numerical simulations distinguish injected signal from fitted congruence and compare noise restricted to the market subspace with noise throughout the text space. At the lowest classifier-agreement scenario, detection remains below $43\%$ even at the largest injected signal. These are conditional checks of the alignment stage, not validation of the text instrument or exclusion bounds on economic effects. The contribution is an auditable non-detection and a specification-sensitive corpus diagnosis, with the inferential limits made explicit.

q-fin.CP↗
arXiv2026

The Limits of Complexity: Why Feature Engineering Beats Deep Learning in Investor Flow Prediction

The application of machine learning to financial prediction has accelerated dramatically, yet the conditions under which complex models outperform simple alternatives remain poorly understood. This paper investigates whether advanced signal processing and deep learning techniques can extract predictive value from investor order flows beyond what simple feature engineering achieves. Using a comprehensive dataset of 2.79 million observations spanning 2,439 Korean equities from 2020--2024, we apply three methodologies: \textit{Independent Component Analysis} (ICA) to recover latent market drivers, \textit{Wavelet Coherence} analysis to characterize multi-scale correlation structure, and \textit{Long Short-Term Memory} (LSTM) networks with attention mechanisms for non-linear prediction. Our results reveal a striking finding: a parsimonious linear model using market capitalization-normalized flows (``Matched Filter'' preprocessing) achieves a Sharpe ratio of 1.30 and cumulative return of 272.6\%, while the full ICA-Wavelet-LSTM pipeline generates a Sharpe ratio of only 0.07 with a cumulative return of $-5.1\%$. The raw LSTM model collapsed to predicting the unconditional mean, achieving a hit rate of 47.5\% -- worse than random. We conclude that in low signal-to-noise financial environments, domain-specific feature engineering yields substantially higher marginal returns than algorithmic complexity. These findings establish important boundary conditions for the application of deep learning to financial prediction.

q-fin.CP↗
arXiv2026

A Geometry-Aware Residual Correction of Hagan's SABR Implied Volatility Formula

This paper proposes a hybrid methodology to improve the approximation of SABR (Stochastic Alpha Beta Rho) implied volatility by combining analytical structure with machine learning. The approach augments the neural-network input representation with geometric features derived from the stochastic differential equations of the SABR model. Unlike approaches that fully replace analytical formulas with black-box models, the proposed framework preserves the analytical backbone of the model. The hybridization operates along two complementary dimensions. First, geometry-aware variables reflecting intrinsic properties of the SABR dynamics are used as structured inputs to the network. Second, the neural network is trained to learn the residual error relative to Hagan's closed-form approximation rather than implied volatility directly. The resulting model acts as a structured residual correction to the analytical formula, retaining interpretability while capturing higher-order effects that are not included in the asymptotic expansion. Numerical experiments conducted over realistic parameter domains, as well as stressed environments, show that the method improves accuracy and robustness compared with both analytical approximations and standard neural-network approaches. Because the correction remains lightweight and structurally consistent with the underlying model, the framework is well suited for real-time pricing and calibration in practical trading environments.

q-fin.CP↗