arXiv subjects

Yutong Yan

Publications and source records attributed to Yutong Yan.

6 recordsLinked to original sources

arXiv2026

Can LLMs Hire Fairly? Racial Bias in Resume Screening

We audit fourteen mainstream large language models (LLMs) for hiring discrimination using the paired-resume methodology of Kline, Rose, and Walters (2022). The sole 2023-vintage model reproduces the pro-White callback gap documented in field experiments on labor market discrimination ($+2.12$ pp, significant at the 1\% level). Every model released in 2024 or after shows either a null gap or a significant pro-Black reversal (up to $-3.01$ pp). The same pattern holds on the gender axis. Based on 24,024 paired postings per model across 14 models, our results document a reversal in the direction of algorithmic hiring bias across model generations.

cs.CL↗

arXiv2026

Debiasing LLMs by Fine-tuning

Prior research shows that large language models (LLMs) exhibit systematic extrapolation bias when forming predictions from both experimental and real-world data, and that prompt-based approaches appear limited in alleviating this bias. We propose a supervised fine-tuning (SFT) approach that uses Low-Rank Adaptation (LoRA) to train off-the-shelf LLMs on instruction datasets constructed from rational benchmark forecasts. By intervening at the parameter level, SFT changes how LLMs map observed information into forecasts and thereby mitigates extrapolation bias. We evaluate the fine-tuned model in two settings: controlled forecasting experiments and cross-sectional stock return prediction. In both settings, fine-tuning corrects the extrapolative bias out-of-sample, establishing a low-cost and generalizable method for debiasing LLMs.

q-fin.GN↗

arXiv2026

DatedGPT: Preventing Lookahead Bias in Large Language Models with Time-Aware Pretraining

Large language models pretrained on internet-scale data risk lookahead bias in forecasting tasks, as they may have already seen the true outcome during training. To address this, we present DatedGPT, a family of twelve 1.3B-parameter language models trained from scratch on approximately 100 billion tokens each with strict annual data cutoffs spanning 2013 to 2024, together with DatedInstruct, an instruction dataset grounded in each year's documents to prevent leakage during post-training. The models are competitive with open models of similar scale, and perplexity-based probing confirms that each model's knowledge is bounded by its cutoff year. On stock return prediction over 61,000 firm-day news headlines, DatedGPT-instruct achieves an annualised Sharpe ratio of $3.20$ under the lookahead-bias-free setup. Lookahead-biased models, whose training data covers the outcome period, add a lookahead premium of $26.4$ b.p. per standard deviation, significant at the 1% level. The series thus enables direct analysis of lookahead bias in financial forecasting. We provide an interactive web demo that allows users to query and compare responses from models across different cutoff years, available at www.datedgpt.com.

cs.CL↗

arXiv2025

Detecting Lookahead Bias in LLM Forecasts

We develop a statistical procedure to detect lookahead bias in economic forecasts generated by large language models (LLMs). Using a date-only recall query for a firm-date pair, we estimate the probability that the LLM has internalized information about the realized outcome, a statistic we term Lookahead Propensity (LAP). LAP is materially positive throughout the in-sample period and collapses essentially to zero right after the training-data cutoff. We show that a positive interaction between LAP and the LLM forecast in an accuracy regression indicates lookahead-bias contamination, and apply the test to two forecasting tasks: news headlines predicting stock returns and earnings call transcripts predicting capital expenditures. In both applications, the LLM forecast's predictive power is amplified on high-LAP firm-date pairs, and the interaction loses significance on post-training-cutoff samples. Our test provides a cost-efficient, diagnostic tool for assessing the validity and reliability of LLM-generated forecasts.

q-fin.GN↗

arXiv2025

The predicable dance of random walk: local limit theorems on finitely-generated abelian groups

In random walk theory, it is customary to assume that a given walk is irreducible and/or aperiodic. While these prevailing assumptions make particularly tractable the analysis of random walks and help to highlight their diffusive nature, they eliminate a natural phenomenon: the dance. This dance can be seen, for example, in the so-called simple random walk on the integers where a random walker moves back and forth between even and odd integers. It can also be seen in the random walk on the integer lattice that takes only those steps available to a knight on a chess board. In this work, we develop a general Fourier-analytic method to describe random walks on finitely-generated abelian groups, making no assumptions concerning aperiodicity or irreducibility. Our main result is a local central limit theorems that describes the large-time behavior of the transition probabilities for any random walk whose driving measure has finite second moments. Generalizing the ubiquitous Gaussian approximation, our asymptotic is given as a product of two functions, one that describes the diffusion and another that describes the dance. Interestingly, our ``dance function" is gotten as a Haar integral over a subgroup of the group's Pontryagin dual. Our results recapture many long-standing results, especially the local limit theorems made famous by F. Spitzer's classic text, Principles of Random Walk.

math.PR↗

arXiv2020

Two-stage multi-scale breast mass segmentation for full mammogram analysis without user intervention

Mammography is the primary imaging modality used for early detection and diagnosis of breast cancer. X-ray mammogram analysis mainly refers to the localization of suspicious regions of interest followed by segmentation, towards further lesion classification into benign versus malignant. Among diverse types of breast abnormalities, masses are the most important clinical findings of breast carcinomas. However, manually segmenting breast masses from native mammograms is time-consuming and error-prone. Therefore, an integrated computer-aided diagnosis system is required to assist clinicians for automatic and precise breast mass delineation. In this work, we present a two-stage multi-scale pipeline that provides accurate mass contours from high-resolution full mammograms. First, we propose an extended deep detector integrating a multi-scale fusion strategy for automated mass localization. Second, a convolutional encoder-decoder network using nested and dense skip connections is employed to fine-delineate candidate masses. Unlike most previous studies based on segmentation from regions, our framework handles mass segmentation from native full mammograms without any user intervention. Trained on INbreast and DDSM-CBIS public datasets, the pipeline achieves an overall average Dice of 80.44% on INbreast test images, outperforming state-of-the-art. Our system shows promising accuracy as an automatic full-image mass segmentation system. Extensive experiments reveals robustness against the diversity of size, shape and appearance of breast masses, towards better interaction-free computer-aided diagnosis.

eess.IV↗