arXiv subjects

Jianfei Xiao

Publications and source records attributed to Jianfei Xiao.

8 recordsLinked to original sources

arXiv2026

From Empathy to Personalized Empathy: Adapting Empathetic Strategies to Individual Users

As Large Language Models (LLMs) are increasingly deployed in long-term interactions with users, empathy has become an increasingly important capability. However, existing research overlooks the influence of users' personality traits on empathetic strategies during long-term interactions. To address this gap, we introduce the task of personalized empathy, which focuses on adapting empathetic strategies according to users' personalized characteristics derived from history. To study and enhance this capability, we construct PersonaEmp, a personalized empathy dataset built from long-term user-AI interactions, featuring rich user histories, persona information, and empathy-seeking queries. We further propose PereGRM, a reward modeling framework that combines the empathy evaluation structure with dynamic evaluation criteria generation for fine-grained reward modeling. Experimental results across different settings and multiple judge models show that PereGRM consistently achieves the strongest performance improvements, indicating its effectiveness for enhancing personalized empathetic capabilities.

cs.CL↗

arXiv2026

Towards Root Memories: Benchmarking and Enhancing Implicit Logical Memory Retrieval for Personalized LLMs

Memory systems are essential for personalized Large Language Models (LLMs). However, existing retrieval methods in these systems primarily rely on semantic similarity, potentially missing logically critical memories with limited semantic overlap. Current benchmarks remain inadequate for evaluating this problem. To address this gap, we construct IMLogic, the first high-quality benchmark targeting implicit logical memory retrieval in long-dialogue scenarios. Motivated by this challenge, we introduce root memory, a structured, decision-preserving representation that distills reusable personalized logic from long-term user histories. We then propose RootMem, a plug-and-play framework that first distills raw histories into structured root memories and then uses an LLM-based router to activate logically relevant ones, complementing semantic retrieval with personalized decision logic. Extensive experiments demonstrate that RootMem significantly outperforms the strongest retrieval baselines and consistently boosts the accuracy of existing memory agents. Our benchmark and codes will be available at https://anonymous.4open.science/r/IMLogic-DBB3.

cs.CL↗

arXiv2026

AlpsBench: An LLM Personalization Benchmark for Real-Dialogue Memorization and Preference Alignment

As Large Language Models (LLMs) evolve into lifelong AI assistants, LLM personalization has become a critical frontier. However, progress is currently bottlenecked by the absence of a gold-standard evaluation benchmark. Existing benchmarks either overlook personalized information management that is critical for personalization or rely heavily on synthetic dialogues, which exhibit an inherent distribution gap from real-world dialogue. To bridge this gap, we introduce AlpsBench, An LLM PerSonalization benchmark derived from real-world human-LLM dialogues. AlpsBench comprises 2,500 long-term interaction sequences curated from WildChat, paired with human-verified structured memories that encapsulate both explicit and implicit personalization signals. We define four pivotal tasks - personalized information extraction, updating, retrieval, and utilization - and establish protocols to evaluate the entire lifecycle of memory management. Our benchmarking of frontier LLMs and memory-centric systems reveals that: (i) models struggle to reliably extract latent user traits; (ii) memory updating faces a performance ceiling even in the strongest models; (iii) retrieval accuracy declines sharply in the presence of large distractor pools; and (iv) while explicit memory mechanisms improve recall, they do not inherently guarantee more preference-aligned or emotionally resonant responses. AlpsBench aims to provide a comprehensive framework.

cs.CL↗

arXiv2025

A non-invasive dry-transfer method for fabricating mesoscopic devices on sensitive materials

Many materials with novel or exotic properties are highly sensitive to environmental factors such as air, solvents, and heat, which complicates device fabrication and limits their potential applications. Here, we present a universal submicron fabrication method for mesoscopic devices using a dry-transfer technique, tailored specifically for sensitive materials. This approach utilizes PMMA masks, combined with a water-dissoluble coating as a sacrificial layer, to ensure that sensitive materials are processed without exposure to harmful environmental conditions. The entire fabrication process is carried out in a glove box, employing dry techniques that avoid air, solvents, and heat exposure, culminating in an encapsulation step. We demonstrate the utility of this method by fabricating and characterizing K2Cr3As3 and WTe2 devices, a one- and two-dimensional material, respectively. The results show that our technique preserves the integrity of the materials, provides excellent contact interfaces, and is broadly applicable to a range of sensitive materials.

cond-mat.mes-hall↗

arXiv2025

Quantized conductance in a CVD-grown nanoribbon with hidden Rashba effect

Quantized conductance in quasi-one-dimensional systems not only provides a hallmark of ballistic transport, but also serves as a gateway for exploring quantum phenomena. Recently, a unique hidden Rashba effect attracts tremendous attention, which arises from the compensation of opposite spin polarizations of a Rashba bilayer in inversion symmetric crystals with dipole fields, such as bismuth oxyselenide ($\mathrm{Bi}_{2}\mathrm{O}_{2}\mathrm{Se}$). However, investigating this effect utilizing conductance quantization is still challenging. Here we report the conductance quantization observed in a chemical vapor deposition (CVD)-grown high-mobility $\mathrm{Bi}_{2}\mathrm{O}_{2}\mathrm{Se}$ nanoribbon, where quantized conductance plateaus up to $44\cdot 2e^{2}/{h}$ ($e$ is the elementary charge, $h$ is the Planck constant, and the factor $2$ results from spin degeneracy) are achieved at zero magnetic field. Due to the hidden Rashba effect, the quantized conductance remains in multiples of $2e^{2}/{h}$ without Zeeman splitting even under magnetic field up to $12$ T. Moreover, within a specific range of magnetic field, the plateau sequence exhibits the Pascal triangle series, namely $(1,3,6,10,15\dots )\cdot 2e^{2}/{h}$, reflecting the interplay of size quantization in two transverse directions. These observations are well captured by an effective hidden Rashba bilayer model. Our results demonstrate $\mathrm{Bi}_{2}\mathrm{O}_{2}\mathrm{Se}$ as a compelling platform for spintronics and the investigation of emergent phenomena.

cond-mat.mes-hall↗

arXiv2025

Diamond-shaped evolution of the superconducting interference pattern in NbTiN weak-link Josephson junctions

The application of in-plane magnetic fields to Josephson junctions enables fundamental exploration of quantum phenomena, including Zeeman-driven 0-$π$ transitions and planar topological superconductivity. However, intrinsic orbital effects arising from nanoscale rippled geometries in practical devices can dominate phase interference signatures, complicating their interpretation. Here, we experimentally probe superconducting interference in NbTiN weak-link Josephson junctions under combined perpendicular and in-plane magnetic fields. The critical supercurrent reveals a distinct diamond-shaped interference pattern, with nodes progressively opening and evolving into V-shaped features, reminiscent of suppression-recovery patterns associated with 0-$π$ transitions. We theoretically analyze the interplay between orbital effects from rippled geometries and non-uniform supercurrent density distributions, demonstrating that their synergistic interaction could reproduce the experimentally observed interference evolution. Our findings elucidate how geometric imperfections and current inhomogeneity cooperatively reshape phase interference, providing critical insights into orbital-dominated phenomena in Josephson systems.

cond-mat.mes-hall↗

arXiv2024

Baichuan2-Sum: Instruction Finetune Baichuan2-7B Model for Dialogue Summarization

Large language models (LLMs) like Llama, Baichuan and Bloom models show remarkable ability with instruction fine-tuning in many natural language tasks. Nevertheless, for the dialogue summarization task, which aims to generate summaries for different roles in dialogue, most of the state-of-the-art methods conduct on small models (e.g Bart and Bert). Existing methods try to add task specified optimization on small models like adding global-local centrality score to models. In this paper, we propose an instruction fine-tuning model: Baichuan2-Sum, for role-oriented diaglouge summarization. By setting different instructions for different roles, the model can learn from the dialogue interactions and output the expected summaries. Furthermore, we applied NEFTune technique to add suitable noise during training to improve the results. The experiments demonstrate that the proposed model achieves the new state-of-the-art results on two public dialogue summarization datasets: CSDS and SAMSUM. We release our model and related codes to facilitate future studies on dialogue summarization task.

cs.CL↗

arXiv2022

Bi2O2Se nanowires presenting high mobility and strong spin-orbit coupling

Systematic electrical transport characterizations were performed on high-quality Bi2O2Se nanowires to illustrate its great transport properties and further application potentials in spintronics. Bi2O2Se nanowires synthesized by chemical vapor deposition method presented a high field-effect mobility up to 1.34*104 cm2V-1s-1, and exhibited ballistic transport in the low back-gate voltage (Vg) regime where conductance plateaus were observed. When further increasing the electron density by increasing Vg, we entered the phase coherent regime and weak antilocalization (WAL) was observed. The spin relaxation length extracted from the WAL was found to be gate tunable, ranging from ~100 nm to ~250 nm and reaching a stronger spin-obit coupling (SOC) than the two-dimensional counterpart (flakes). We attribute the strong SOC and the gate tunability to the presence of a surface accumulation layer which induces a strong inversion asymmetry on the surface. Such scenario was supported by the observation of two Shubnikov-de Haas oscillation frequencies that correspond to two types of carriers, one on the surface, and the other in the bulk. The high-quality Bi2O2Se nanowires with a high mobility and a strong SOC can act as a very prospective material in future spintronics.

cond-mat.mes-hall↗