arXiv subjects

Tracy Xiao Liu

Publications and source records attributed to Tracy Xiao Liu.

8 recordsLinked to original sources

arXiv2026

Understanding the Mechanism of Altruism in Large Language Models

Altruism is fundamental to human societies, fostering cooperation and social cohesion. Recent studies suggest that large language models (LLMs) can display human-like prosocial behavior, but the internal computations that produce such behavior remain poorly understood. We investigate the mechanisms underlying LLM altruism using sparse autoencoders (SAEs). In a standard Dictator Game, minimal-pair prompts that differ only in social stance (generous versus selfish) induce large, economically meaningful shifts in allocations. Leveraging this contrast, we identify a set of SAE features (0.024% of all features across the model's layers) whose activations are strongly associated with the behavioral shift. To interpret these features, we use benchmark tasks motivated by dual-process theories to classify a subset as primarily heuristic (System 1) or primarily deliberative (System 2). Causal interventions validate their functional role: activation patching and continuous steering of this feature direction reliably shift allocation distributions, with System 2 features exerting a more proximal influence on the model's final output than System 1 features. The same steering direction generalizes across multiple social-preference games. Together, these results enhance our understanding of artificial cognition by translating altruistic behaviors into identifiable network states and provide a framework for aligning LLM behavior with human values, thereby informing more transparent and value-aligned deployment.

econ.GN↗

arXiv2026

Not All Students Engage Alike: Multi-Institution Patterns in GenAI Tutor Use

The emergence of generative artificial intelligence (GenAI) has created unprecedented opportunities to provide individualized learning support in classrooms as automated tutoring systems at scale. However, concerns have been raised that students may engage with these tools in ways that do not support learning. Moreover, student engagement with GenAI Tutors may vary across instructional contexts, potentially leading to unequal learning experiences. In this study, we utilize de-identified student interaction logs from an existing GenAI Tutor and the learning management system in which it is embedded. We systematically examined student engagement (N = 11,406) with the tool across 200 classes in ten post-secondary institutions through a two-stage pipeline: First, we identified four distinct engagement types at the conversation session level. In particular, 10.4% of them were "shallow engagement" where copy-pasting behavior was prevalent. Then, at the student level, we show that students transitioned across engagement types over time. However, students who exhibited shallow engagement with the tool were more likely to remain in this mode, whereas those who engaged deeply with the tool transitioned more flexibly across engagement types. Finally, at both the session and student levels, we show substantial heterogeneity in student engagement across institution selectivity and course disciplines. In particular, students from highly selective institutions were more likely to exhibit deep engagement. Together, our study advances the understanding of how GenAI Tutors are used in authentic educational settings and provides a framework for analyzing student engagement with GenAI Tutors, with implications for responsible implementation at scale.

cs.CY↗

arXiv2025

How General Are Measures of Choice Consistency? Evidence from Experimental and Scanner Data

Choice consistency with utility maximization, as a key assumption in economics, has been extensively used to evaluate decision quality of individuals and to predict real-world outcomes across different contexts. Here we investigate the generalizability of consistency measures derived from budgetary decisions in the lab-in-the-field experiment and purchasing decisions using supermarket scanner data. In the first study, we observe a lack of correlation between consistency scores derived from risky decisions in the experiment and those from supermarket food purchasing decisions. In the second study, we observe moderate correlations between experimental tasks and low to moderate correlations across purchasing categories and over time periods within the supermarket. Moreover, consistency in the two settings exhibits distinct predictive validity in predicting consumer behavior. These results suggest that choice consistency, as a measure of decision quality, may be better characterized as a multidimensional skill set rather than a single-dimensional ability.

econ.GN↗

arXiv2025

When Experimental Economics Meets Large Language Models: Evidence-based Tactics

Advancements in large language models (LLMs) have sparked a growing interest in measuring and understanding their behavior through experimental economics. However, there is still a lack of established guidelines for designing economic experiments for LLMs. Inspired by principles from experimental economics with insights from LLM research in artificial intelligence, we outline key considerations in the experimental design and implementation stage, and perform two sets of experiments to assess the impact of these considerations on LLMs' responses. Based on our findings, we discuss seven practical tactics for conducting experiments with LLMs. Our study enhances the design, replicability, and generalizability of LLM experiments, and broadens the scope of experimental economics in the digital age.

econ.GN↗

arXiv2023

The Emergence of Economic Rationality of GPT

As large language models (LLMs) like GPT become increasingly prevalent, it is essential that we assess their capabilities beyond language processing. This paper examines the economic rationality of GPT by instructing it to make budgetary decisions in four domains: risk, time, social, and food preferences. We measure economic rationality by assessing the consistency of GPT's decisions with utility maximization in classic revealed preference theory. We find that GPT's decisions are largely rational in each domain and demonstrate higher rationality score than those of human subjects in a parallel experiment and in the literature. Moreover, the estimated preference parameters of GPT are slightly different from human subjects and exhibit a lower degree of heterogeneity. We also find that the rationality scores are robust to the degree of randomness and demographic settings such as age and gender, but are sensitive to contexts based on the language frames of the choice situations. These results suggest the potential of LLMs to make good decisions and the need to further understand their capabilities, limitations, and underlying mechanisms.

econ.GN↗

arXiv2021

BONUS! Maximizing Surprise

Multi-round competitions often double or triple the points awarded in the final round, calling it a bonus, to maximize spectators' excitement. In a two-player competition with $n$ rounds, we aim to derive the optimal bonus size to maximize the audience's overall expected surprise (as defined in [7]). We model the audience's prior belief over the two players' ability levels as a beta distribution. Using a novel analysis that clarifies and simplifies the computation, we find that the optimal bonus depends greatly upon the prior belief and obtain solutions of various forms for both the case of a finite number of rounds and the asymptotic case. In an interesting special case, we show that the optimal bonus approximately and asymptotically equals to the "expected lead", the number of points the weaker player will need to come back in expectation. Moreover, we observe that priors with a higher skewness lead to a higher optimal bonus size, and in the symmetric case, priors with a higher uncertainty also lead to a higher optimal bonus size. This matches our intuition since a highly asymmetric prior leads to a high "expected lead", and a highly uncertain symmetric prior often leads to a lopsided game, which again benefits from a larger bonus.

cs.GT↗

arXiv2021

SURPRISE! and When to Schedule It

Information flow measures, over the duration of a game, the audience's belief of who will win, and thus can reflect the amount of surprise in a game. To quantify the relationship between information flow and audiences' perceived quality, we conduct a case study where subjects watch one of the world's biggest esports events, LOL S10. In addition to eliciting information flow, we also ask subjects to report their rating for each game. We find that the amount of surprise in the end of the game plays a dominant role in predicting the rating. This suggests the importance of incorporating when the surprise occurs, in addition to the amount of surprise, in perceived quality models. For content providers, it implies that everything else being equal, it is better for twists to be more likely to happen toward the end of a show rather than uniformly throughout.

cs.MA↗

arXiv2017

Online Red Packets: A Large-scale Empirical Study of Gift Giving on WeChat

Gift giving is a ubiquitous social phenomenon, and red packets have been used as monetary gifts in Asian countries for thousands of years. In recent years, online red packets have become widespread in China through the WeChat platform. Exploiting a unique dataset consisting of 61 million group red packets and seven million users, we conduct a large-scale, data-driven study to understand the spread of red packets and the effect of red packets on group activity. We find that the cash flows between provinces are largely consistent with provincial GDP rankings, e.g., red packets are sent from users in the south to those in the north. By distinguishing spontaneous from reciprocal red packets, we reveal the behavioral patterns in sending red packets: males, seniors, and people with more in-group friends are more inclined to spontaneously send red packets, while red packets from females, youths, and people with less in-group friends are more reciprocal. Furthermore, we use propensity score matching to study the external effects of red packets on group dynamics. We show that red packets increase group participation and strengthen in-group relationships, which partly explain the benefits and motivations for sending red packets.

cs.SI↗