115 citations · 450 across the 154 of their papers we have counts for
22 papers · 1 filter
MetaSteer: Context-Conditioned, nonlinear Steering via Attention-Projection Adaptation
Mehdi Jafari, Hao Xue, Flora Salim
Steering large language models typically relies on linear, context-independent interventions in activation space, an assumption that recent work has challenged and that can induce…
Scope3Trace: Evidence-Based Identification and Extraction of Scope 3 GHG Emissions from Sustainability Reports
Siyuan Zheng, Yifan Duan, Chao Xue +1
Scope 3 greenhouse gas (GHG) emissions account for the majority of corporate carbon footprints, yet remain difficult to analyze at scale due to sparse disclosures, heterogeneous re…
Group Alignment-Induced Sycophancy: A Two-Sided Evaluation of Steerable Pluralistic Alignment
Haokai Zhao, Yunze Xiao, Weihao Xuan +3
Group alignment adapts a language model to a demographic group to produce responses that reflect the group's opinions, values, and preferences. Sycophancy, a well-documented by-pro…
Evaluating LLM Personalization via Semantic Constraint Verification
Xuran Li, Guanqin Zhang, Imran Razzak +4
Current evaluation paradigms for Large Language Model (LLM) personalization rely heavily on brittle surface-matching metrics or computationally expensive LLM-as-a-judge protocols,…
Didact: A Cross-Domain Capability Discovery System for Defence
Aarya Bodhankar, Aditya Joshi, Bao Gia Doan +3
Policymakers in defence and defence-aligned sectors must monitor rapidly evolving research alongside sector priorities relevant to operational and strategic needs. In practice, the…
A Benchmark Construction and Evaluation Framework for Specialist Domains: Case Study on Defense-related Documents
Bao Gia Doan, Aditya Joshi, Pantelis Elinas +4
RAG-based question-answering (QA) in specialist domains faces a cold-start problem: lack of evaluative benchmarks and absence of labeled data for post-training. We present DoRA (Do…
Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty
Chao Xue, Yao Wang, Mengqiao Liu +11
Recent advancements in the Generative Reward Model (GRM) have demonstrated its potential to enhance the reasoning abilities of LLMs through Chain-of-Thought (CoT) prompting. Despit…
Why Supervised Fine-Tuning Fails to Learn: A Systematic Study of Incomplete Learning in Large Language Models
Chao Xue, Yao Wang, Mengqiao Liu +11
Supervised Fine-Tuning (SFT) is the standard approach for adapting large language models (LLMs) to downstream tasks. However, we observe a persistent failure mode: even after conve…
Mechanistic Indicators of Steering Effectiveness in Large Language Models
Mehdi Jafari, Hao Xue, Flora Salim
Activation-based steering enables Large Language Models (LLMs) to exhibit targeted behaviors by intervening on intermediate activations without retraining. Despite its widespread u…
Is my model "mind blurting"? Interpreting the dynamics of reasoning tokens with Recurrence Quantification Analysis (RQA)
Quoc Tuan Pham, Mehdi Jafari, Flora Salim
Test-time compute is central to large reasoning models, yet analysing their reasoning behaviour through generated text is increasingly impractical and unreliable. Response length i…
RELOOP: Recursive Retrieval with Multi-Hop Reasoner and Planners for Heterogeneous QA
Ruiyi Yang, Hao Xue, Imran Razzak +2
Retrieval-augmented generation (RAG) remains brittle on multi-step questions and heterogeneous evidence sources, trading accuracy against latency and token/tool budgets. This paper…
Alternatives To Next Token Prediction In Text Generation -- A Survey
Charlie Wyatt, Aditya Joshi, Flora Salim
The paradigm of Next Token Prediction (NTP) has driven the unprecedented success of Large Language Models (LLMs), but is also the source of their most persistent weaknesses such as…
What am I missing here?: Evaluating Large Language Models for Masked Sentence Prediction
Charlie Wyatt, Aditya Joshi, Flora Salim
Transformer-based models primarily rely on Next Token Prediction (NTP), which predicts the next token in a sequence based on the preceding context. However, NTP's focus on single-t…
ZARA: Training-Free Motion Time-Series Reasoning via Evidence-Grounded LLM Agents
Zechen Li, Baiyu Chen, Hao Xue +1
Motion sensor time-series are central to Human Activity Recognition (HAR), yet conventional approaches are constrained to fixed activity sets and typically require costly parameter…
Multi-Stage Verification-Centric Framework for Mitigating Hallucination in Multi-Modal RAG
Baiyu Chen, Wilson Wongso, Xiaoqian Hu +2
This paper presents the technical solution developed by team CRUISE for the KDD Cup 2025 Meta Comprehensive RAG Benchmark for Multi-modal, Multi-turn (CRAG-MM) challenge. The chall…
Evaluating the Bias in LLMs for Surveying Opinion and Decision Making in Healthcare
Yonchanok Khaokaew, Flora D. Salim, Andreas Züfle +5
Generative agents have been increasingly used to simulate human behaviour in silico, driven by large language models (LLMs). These simulacra serve as sandboxes for studying human b…
RIDE: Enhancing Large Language Model Alignment through Restyled In-Context Learning Demonstration Exemplars
Yuncheng Hua, Lizhen Qu, Zhuang Li +3
Alignment tuning is crucial for ensuring large language models (LLMs) behave ethically and helpfully. Current alignment approaches require high-quality annotations and significant…
Harnessing Test-time Adaptation for NLU tasks Involving Dialects of English
Duke Nguyen, Aditya Joshi, Flora Salim
Test-time domain adaptation (TTDA) is an excellent method which helps generalize models across domains, tasks, and distributions without the use of labeled datasets. Thus, TTDA is…
Enhancing Conversational Agents with Theory of Mind: Aligning Beliefs, Desires, and Intentions for Human-Like Interaction
Mehdi Jafari, Devin Yuncheng Hua, Hao Xue +1
Natural language interaction with agentic Artificial Intelligence (AI), driven by Large Language Models (LLMs), is expected to remain a dominant paradigm in the near future. While…
SensorLLM: Aligning Large Language Models with Motion Sensors for Human Activity Recognition
Zechen Li, Shohreh Deldari, Linyao Chen +2
We introduce SensorLLM, a two-stage framework that enables Large Language Models (LLMs) to perform human activity recognition (HAR) from sensor time-series data. Despite their stro…
Measuring Misogyny in Natural Language Generation: Preliminary Results from a Case Study on two Reddit Communities
Aaron J. Snoswell, Lucinda Nelson, Hao Xue +3
Generic `toxicity' classifiers continue to be used for evaluating the potential for harm in natural language generation, despite mounting evidence of their shortcomings. We conside…
Human Mobility Question Answering (Vision Paper)
Hao Xue, Flora D. Salim
Question answering (QA) systems have attracted much attention from the artificial intelligence community as they can learn to answer questions based on the given knowledge source (…