arXiv subjects

Shan Dong

Publications and source records attributed to Shan Dong.

8 recordsLinked to original sources

arXiv2026

On Reasoning Behind Next Occupation Recommendation

In this work, we develop a novel reasoning approach to enhance the performance of large language models (LLMs) in future occupation prediction. In this approach, a reason generator first derives a ``reason'' for a user using his/her past education and career history. The reason summarizes the user's preference and is used as the input of an occupation predictor to recommend the user's next occupation. This two-step occupation prediction approach is, however, non-trivial as LLMs are not aligned with career paths or the unobserved reasons behind each occupation decision. We therefore propose to fine-tune LLMs improving their reasoning and occupation prediction performance. We first derive high-quality oracle reasons, as measured by factuality, coherence and utility criteria, using a LLM-as-a-Judge. These oracle reasons are then used to fine-tune small LLMs to perform reason generation and next occupation prediction. Our extensive experiments show that: (a) our approach effectively enhances LLM's accuracy in next occupation prediction making them comparable to fully supervised methods and outperforming unsupervised methods; (b) a single LLM fine-tuned to perform reason generation and occupation prediction outperforms two LLMs fine-tuned to perform the tasks separately; and (c) the next occupation prediction accuracy depends on the quality of generated reasons. Our code is available at https://github.com/Sarasarahhhhh/job_prediction.

cs.CL↗

arXiv2025

Effects of Theory of Mind and Prosocial Beliefs on Steering Human-Aligned Behaviors of LLMs in Ultimatum Games

Large Language Models (LLMs) have shown potential in simulating human behaviors and performing theory-of-mind (ToM) reasoning, crucial for complex social interactions. We investigate ToM reasoning's role in aligning agentic behaviors with human norms in negotiation tasks, using the ultimatum game as our referenced task. We initialized LLM agents with different prosocial beliefs (Greedy, Fair, Selfless) and reasoning methods (chain of thought and ToM reasoning of varying levels), examining their decision-making process and outcome across multiple LLMs, including reasoning models like o3-mini and DeepSeek-R1 Distilled Qwen 32B. We perform 2,700 simulations to show that ToM reasoning enhances behavioral alignment with human, decision-making consistency, and negotiation outcomes. Consistent with prior findings, reasoning LLMs exhibit limited capability compared to ToM-enhanced LLMs, with different game roles benefiting from different ToM orders. Fair proposers and responders accepting offers were the most consistent with their strategic reasonings, whereas all agents showed strong consistencies with human beliefs when rejecting offers, except when the offer was fair. Human verification further revealed that Llama 3.3 70B produces reasoning most consistent with its actions and beliefs. Our findings advance understanding of ToM's role in human-AI interaction and cooperative decision-making. The code used for our experiments can be found at https://github.com/Stealth-py/UltimatumToM.

cs.CL↗

arXiv2024

MathHay: An Automated Benchmark for Long-Context Mathematical Reasoning in LLMs

Recent large language models (LLMs) have demonstrated versatile capabilities in long-context scenarios. Although some recent benchmarks have been developed to evaluate the long-context capabilities of LLMs, there is a lack of benchmarks evaluating the mathematical reasoning abilities of LLMs over long contexts, which is crucial for LLMs' application in real-world scenarios. In this paper, we introduce MathHay, an automated benchmark designed to assess the long-context mathematical reasoning capabilities of LLMs. Unlike previous benchmarks like Needle in a Haystack, which focus primarily on information retrieval within long texts, MathHay demands models with both information-seeking and complex mathematical reasoning abilities. We conduct extensive experiments on MathHay to assess the long-context mathematical reasoning abilities of eight top-performing LLMs. Even the best-performing model, Gemini-1.5-Pro-002, still struggles with mathematical reasoning over long contexts, achieving only 51.26% accuracy at 128K tokens. This highlights the significant room for improvement on the MathHay benchmark.

cs.CL↗

arXiv2024

All in an Aggregated Image for In-Image Learning

This paper introduces a new in-context learning (ICL) mechanism called In-Image Learning (I$^2$L) that combines demonstration examples, visual cues, and chain-of-thought reasoning into an aggregated image to enhance the capabilities of Large Multimodal Models (e.g., GPT-4V) in multimodal reasoning tasks. Unlike previous approaches that rely on converting images to text or incorporating visual input into language models, I$^2$L consolidates all information into an aggregated image and leverages image processing, understanding, and reasoning abilities. This has several advantages: it reduces inaccurate textual descriptions of complex images, provides flexibility in positioning demonstration examples, and avoids multiple input images and lengthy prompts. We also introduce I$^2$L-Hybrid, a method that combines the strengths of I$^2$L with other ICL methods. Specifically, it uses an automatic strategy to select the most suitable method (I$^2$L or another certain ICL method) for a specific task instance. We conduct extensive experiments to assess the effectiveness of I$^2$L and I$^2$L-Hybrid on MathVista, which covers a variety of complex multimodal reasoning tasks. Additionally, we investigate the influence of image resolution, the number of demonstration examples in a single image, and the positions of these demonstrations in the aggregated image on the effectiveness of I$^2$L. Our code is publicly available at https://github.com/AGI-Edgerunners/IIL.

cs.CV↗

arXiv2023

Robust high-temperature topological excitonic insulator of transition-metal carbides (MXenes)

Topological excitonic insulators combine topological edge states and spontaneous exciton condensation, with dual functionality of topological insulators and excitonic insulators. Yet, they are very rare and little is known about their formation. In this work, we find that a mechanism dubbed as parity frustration prevents excitonic instability in usual topological insulators, and those whose band inversion is independent of spin-orbit coupling are possible candidates. We verify this by first-principles calculations on monolayer transition-metal carbides (MXenes), which show a robust thermal-equilibrium exciton condensation, being sufficient for topological applications at room temperature. Such a state can be identified by angle-resolved photoemission spectroscopy and transport measurement. Our work provides not only a guide for finding more topological excitonic insulators, but also a new platform for studying the interplay between non-trivial band topology and quantum many-body effects.

cond-mat.mes-hall↗

arXiv2021

Excitonic Instability and Electronic Property of Two-dimensional AlSb Limit

Motivated by the recent synthesis of two-dimensional monolayer AlSb, we theoretically investigate its ground state and electronic properties using the first-principles calculations coupled with Bethe-Salpeter equation. An excitonic instability is revealed as a result of larger exciton binding energy than the corresponding one-electron energy gap by $\sim$0.1 eV, which is an indicative of a many-body ground state accompanied by spontaneous exciton generation. Including the spin-orbit coupling is proven to be a must to correctly predict the ground state. At room temperature, the two-dimensional monolayer AlSb is a direct gap semiconductor with phonon-limited electron and hole mobilities both around 1700 cm$^2$/V$\cdot$s. These results show that monolayer AlSb may provide a promising platform for realization of the excitonic insulator and for applications in the next-generation electronic devices.

cond-mat.mtrl-sci↗

arXiv2020

Transition from band insulator to excitonic insulator via alloying Se into Monolayer TiS$_3$: A Computational Study

First-principles density functional theory plus Bethe-Salpeter equation calculations are employed to investigate the electronic and excitonic properties of monolayer titanium trichalcogenide alloys TiS$_{3-x}$Se$_x$ ($x$=1 and 2). It is found that bandgap and exciton binding energy display asymmetric dependence on the substitution of Se for S. While the bandgap can be significantly decreased as compared to that of pristine TiS$_3$, the exciton binding energy just varies a little, regardless of position and concentration of the Se substitution. A negative exciton formation energy is found when the central S atoms are replaced by Se atoms, suggesting a many-body ground state with the spontaneous exciton condensation. Our work thus offers a new insight for engineering an excitonic insulator.

cond-mat.mtrl-sci↗

arXiv2015

Ultralow-frequency collective compression mode and strong interlayer coupling in multilayer black phosphorus

The recent renaissance of black phosphorus (BP) as a two-dimensional 2D layered material has generated tremendous interest in its tunable electronic band gap and highly anisotropic transport properties that offer new opportunities for device applications. Many of these outstanding properties are attributed to its unique structural characters that still need elucidation. Here we show Raman measurements that reveal an ultralow-frequency collective compression mode (CCM), which is unprecedented among similar 2D layered materials. This novel CCM indicates an unusually strong interlayer coupling in BP, which is quantitatively supported by a phonon frequency analysis and first-principles calculations. Moreover, the CCM and another branch of low-frequency Raman modes shift sensitively with changing number of layers, allowing an accurate determination of the thickness up to tens of atomic layers, which is considerably higher than those previously achieved by using high-frequency Raman modes. These results offer fundamental insights and practical tools for exploring multilayer BP in new device applications.

cond-mat.mes-hall↗