arXiv subjects

Qiqi Jiang

Publications and source records attributed to Qiqi Jiang.

4 recordsLinked to original sources

arXiv2025

A Search for High-Energy Emission from the Remnant of Supernova 1181

Over the previous millennium, only five Galactic supernovae were observed and recorded by contemporary astronomers, and their current-day counterparts subsequently identified. The remnants of four of these have all been very deeply studied, and ultimately detected, by TeV instruments after exposures of typically hundreds of hours. The measured TeV fluxes range from 1 Crab (by definition) down to 0.3% Crab. The location of the fifth supernova remnant tied to a historical record of its supernova (SN 1181) has never been studied at TeV energies. The reason for this is simple - the associated remnant was only identified as such in 2021. The remnant, Pa 30, is an unusual object whose properties are best explained as resulting from a Type Iax supernova explosion. These are a rare sub-type of Type Ia supernovae in which the merging white dwarfs are not fully destroyed by the supernova explosion, leading to a double-degenerate merger product colorfully described as a 'zombie star'. We will present the results of a search for TeV gamma-ray emission from Pa 30 with VERITAS.

astro-ph.HE↗

arXiv2025

ConceptViz: A Visual Analytics Approach for Exploring Concepts in Large Language Models

Large language models (LLMs) have achieved remarkable performance across a wide range of natural language tasks. Understanding how LLMs internally represent knowledge remains a significant challenge. Despite Sparse Autoencoders (SAEs) have emerged as a promising technique for extracting interpretable features from LLMs, SAE features do not inherently align with human-understandable concepts, making their interpretation cumbersome and labor-intensive. To bridge the gap between SAE features and human concepts, we present ConceptViz, a visual analytics system designed for exploring concepts in LLMs. ConceptViz implements a novel dentification => Interpretation => Validation pipeline, enabling users to query SAEs using concepts of interest, interactively explore concept-to-feature alignments, and validate the correspondences through model behavior verification. We demonstrate the effectiveness of ConceptViz through two usage scenarios and a user study. Our results show that ConceptViz enhances interpretability research by streamlining the discovery and validation of meaningful concept representations in LLMs, ultimately aiding researchers in building more accurate mental models of LLM features. Our code and user guide are publicly available at https://github.com/Happy-Hippo209/ConceptViz.

cs.CL↗

arXiv2025

CSRM-LLM: Embracing Multilingual LLMs for Cold-Start Relevance Matching in Emerging E-commerce Markets

As global e-commerce platforms continue to expand, companies are entering new markets where they encounter cold-start challenges due to limited human labels and user behaviors. In this paper, we share our experiences in Coupang to provide a competitive cold-start performance of relevance matching for emerging e-commerce markets. Specifically, we present a Cold-Start Relevance Matching (CSRM) framework, utilizing a multilingual Large Language Model (LLM) to address three challenges: (1) activating cross-lingual transfer learning abilities of LLMs through machine translation tasks; (2) enhancing query understanding and incorporating e-commerce knowledge by retrieval-based query augmentation; (3) mitigating the impact of training label errors through a multi-round self-distillation training strategy. Our experiments demonstrate the effectiveness of CSRM-LLM and the proposed techniques, resulting in successful real-world deployment and significant online gains, with a 45.8% reduction in defect ratio and a 0.866% uplift in session purchase rate.

cs.IR↗

arXiv2025

Dispersion relation for the linear theory of relativistic Rayleigh Taylor instability in magnetized medium revisited

The Rayleigh-Taylor instability (RTI) arises at the interface between two fluids of different densities, notably when a heavier fluid lies above a lighter one in an effective gravitational field. In astrophysical systems with high velocities, relativistic corrections are necessary. We investigate the linear theory of relativistic Rayleigh-Taylor instability (R-RTI) in a magnetized medium, where fluids can move parallel to the interface at relativistic velocities. We chose an "intermediate frame" where fluids on each side of the interface move in opposite directions with identical Lorentz factors gamma_*. This symmetry facilitates analytical derivations and the study of relativistic effects on the instability's dynamics. We derive the correct version of the R-RTI. We find that the instability is activated when the Atwood number A = (rho1 h1 - rho2 h2) / (rho1 h1 + rho2 h2) > 0, where rho1 and rho2 are densities measured in the rest frame of the fluids, and that this criterion does not contain relativistic corrections. The relativistic effect is mostly contained in the Lorentz transformation of the gravitational acceleration g' = g / (gamma_*)^2, which, combined with time dilation, leads to a much slower growth of instability (omega' = omega0 / gamma_*), and a slightly elongated length of the unstable patch, due to weaker g in the intermediate frame. Taking time dilation into account, when viewed in the rest frame of the medium, we expect the instability to grow at a much reduced rate. The analytical results should guide further explorations of instability in systems such as microquasars (muQSOs), active galactic nuclei (AGNs), gamma-ray bursts (GRBs), and radio pulsars (PSRs), where the apparent stability of the jet can be attributed to either the intrinsic stability (e.g. the Atwood number) or the much prolonged duration through which R-RTI can grow.

astro-ph.HE↗