43 citations · 101 across the 29 of their papers we have counts for
35 papers
SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding Agents
Do Xuan Trong, Duc Anh Nguyen, Do Xuan Long +12
We present SWE-MISBEHAVIOR, the first taxonomy and evaluation framework for diagnosing boundary-violating misbehaviors in coding agents. It contains 7.4K human-in-the-loop synthesi…
Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training
Shrey Pandit, Xuan-Phi Nguyen, Yiran Zhao +1
Training a Mixture-of-Experts (MoE) model at long context or large batch size fails as soon as any one component's peak allocation exceeds device memory, so the target is every pea…
EvoHarnessBench: Can Your Agents Keep Pace with an Evolving Harness?
Zixuan Ke, Vaidehi Patil, Haizhou Shi +9
Modern LLM-based agents operate through a harness of tools, reusable skills, and specialist agents that shapes what they observe and what they can do. In practice, this harness con…
Learning Generalizable Behaviors for Terminal Agents
Yihang Yao, Bo Pang, Xuan Phi Nguyen +3
Terminal agents are a compelling application of large language models (LLMs), with the potential to integrate deeply into users' daily workflows. Reinforcement learning (RL) is a k…
Privileged Likelihood Is Not Automatically Value: Three Checks for Token Credit in On-Policy Self-Distillation
Xuan-Phi Nguyen, Zeyu Leo Liu, Yang Li +4
On-policy self-distillation aims to improve upon reinforcement learning from verifiable rewards (RLVR) by providing token-level scores derived from privileged information, such as…
Mixture-of-Parallelisms: Towards Memory-Efficient Training Stack for Mixture-of-Experts Models
Xuan-Phi Nguyen, Shrey Pandit, Yiran Zhao +3
This paper showcases a memory-efficient training stack for Mixture-of-Experts (MoE) models. It is a training paradigm that combines and specializes various existing and novel paral…
Least-Loaded Expert Parallelism: Load Balancing An Imbalanced Mixture-of-Experts
Xuan-Phi Nguyen, Shrey Pandit, Austin Xu +2
Mixture-of-Experts (MoE) models are typically pre-trained with explicit load-balancing constraints to ensure statistically balanced expert routing. Despite this, we observe that ev…
MAS-Orchestra: Understanding and Improving Multi-Agent Reasoning Through Holistic Orchestration and Controlled Benchmarks
Zixuan Ke, Yifei Ming, Austin Xu +7
While multi-agent systems (MAS) promise elevated intelligence through coordination of agents, current approaches to automatic MAS design under-deliver. Such shortcomings stem from…
Foundational Automatic Evaluators: Scaling Multi-Task Generative Evaluator Training for Reasoning-Centric Domains
Austin Xu, Xuan-Phi Nguyen, Yilun Zhou +3
Finetuning specialized generative evaluators has emerged as a popular paradigm to meet the increasing demand for scalable evaluation during both training and test-time. However, re…
Synthesizing Agentic Data for Web Agents with Progressive Difficulty Enhancement Mechanisms
Shrey Pandit, Xuan-Phi Nguyen, Yifei Ming +4
Web-based 'deep research' agents aim to solve complex question - answering tasks through long-horizon interactions with online tools. These tasks remain challenging, as the underly…
Hard2Verify: A Step-Level Verification Benchmark for Open-Ended Frontier Math
Shrey Pandit, Austin Xu, Xuan-Phi Nguyen +3
Large language model (LLM)-based reasoning systems have recently achieved gold medal-level performance in the IMO 2025 competition, writing mathematical proofs where, to receive fu…
SFR-DeepResearch: Towards Effective Reinforcement Learning for Autonomously Reasoning Single Agents
Xuan-Phi Nguyen, Shrey Pandit, Revanth Gangi Reddy +4
Equipping large language models (LLMs) with complex, interleaved reasoning and tool-use capabilities has become a key focus in agentic AI research, especially with recent advances…
J4R: Learning to Judge with Equivalent Initial State Group Relative Policy Optimization
Austin Xu, Yilun Zhou, Xuan-Phi Nguyen +2
To keep pace with the increasing pace of large language models (LLM) development, model output evaluation has transitioned away from time-consuming human evaluation to automatic ev…
Helpful Agent Meets Deceptive Judge: Understanding Vulnerabilities in Agentic Workflows
Yifei Ming, Zixuan Ke, Xuan-Phi Nguyen +2
Agentic workflows -- where multiple large language model (LLM) instances interact to solve tasks -- are increasingly built on feedback mechanisms, where one model evaluates and cri…
MAS-ZERO: Designing Multi-Agent Systems with Zero Supervision
Zixuan Ke, Austin Xu, Yifei Ming +4
Multi-agent systems (MAS) leveraging the impressive capabilities of Large Language Models (LLMs) hold significant potential for tackling complex tasks. However, most current MAS de…
SweRank: Software Issue Localization with Code Ranking
Revanth Gangi Reddy, Tarun Suresh, JaeHyeok Doo +7
Software issue localization, the task of identifying the precise code locations (files, classes, or functions) relevant to a natural language issue description (e.g., bug report, f…
A Survey of Frontiers in LLM Reasoning: Inference Scaling, Learning to Reason, and Agentic Systems
Zixuan Ke, Fangkai Jiao, Yifei Ming +9
Reasoning is a fundamental cognitive process that enables logical inference, problem-solving, and decision-making. With the rapid advancement of large language models (LLMs), reaso…
Demystifying Domain-adaptive Post-training for Financial LLMs
Zixuan Ke, Yifei Ming, Xuan-Phi Nguyen +2
Domain-adaptive post-training of large language models (LLMs) has emerged as a promising approach for specialized domains such as medicine and finance. However, significant challen…
StructTest: Benchmarking LLMs' Reasoning through Compositional Structured Outputs
Hailin Chen, Fangkai Jiao, Mathieu Ravaut +8
The rapid advancement of large language models (LLMs) demands robust, unbiased, and scalable evaluation methods. However, human annotations are costly to scale, model-based evaluat…
Seed-Free Synthetic Data Generation Framework for Instruction-Tuning LLMs: A Case Study in Thai
Parinthapat Pengpun, Can Udomcharoenchaikit, Weerayut Buaphet +1
We present a synthetic data approach for instruction-tuning large language models (LLMs) for low-resource languages in a data-efficient manner, specifically focusing on Thai. We id…
FaithEval: Can Your Language Model Stay Faithful to Context, Even If "The Moon is Made of Marshmallows"
Yifei Ming, Senthil Purushwalkam, Shrey Pandit +4
Ensuring faithfulness to context in large language models (LLMs) and retrieval-augmented generation (RAG) systems is crucial for reliable deployment in real-world applications, as…
Discovering the Gems in Early Layers: Accelerating Long-Context LLMs with 1000x Input Token Reduction
Zhenmei Shi, Yifei Ming, Xuan-Phi Nguyen +2
Large Language Models (LLMs) have demonstrated remarkable capabilities in handling long context inputs, but this comes at the cost of increased computational resources and latency.…
SFR-RAG: Towards Contextually Faithful LLMs
Xuan-Phi Nguyen, Shrey Pandit, Senthil Purushwalkam +7
Retrieval Augmented Generation (RAG), a paradigm that integrates external contextual information with large language models (LLMs) to enhance factual accuracy and relevance, has em…
ParaICL: Towards Parallel In-Context Learning
Xingxuan Li, Xuan-Phi Nguyen, Shafiq Joty +1
Large language models (LLMs) have become the norm in natural language processing (NLP), excelling in few-shot in-context learning (ICL) with their remarkable abilities. Nonetheless…