arXiv subjects

Ruihong Huang

Publications and source records attributed to Ruihong Huang.

At least 19 recordsLinked to original sources

arXiv2026

Measuring LLM Sycophancy under Sustained Multi-Turn Pressure

Large language models (LLMs) may abandon correct positions when users push back, exhibiting a failure mode known as sycophancy. Existing evaluations typically use short, pre-specified conversations and may therefore miss failures that emerge under sustained, adaptive disagreement. We introduce SPINE, a benchmark in which an LLM proxy plays a persistent but mistaken user and adaptively challenges a target model for up to 25 turns. We evaluate four production systems and three Olmo3-7b variants on 100 false-presupposition and 100 unethical-query items. Our experimental results show that collapse rates increase with conversation length for every model, short-horizon protocols underestimate sycophancy and resistance under sustained pressure remains unreliable across current models. By analyzing models with accessible reasoning traces, we surprisingly found that the correct position often remains represented in a reasoning trace when the response concedes, suggesting that the model chooses to please a user and sycophancy is not due to lack of knowledge or ignorance. Ablations show that adaptive LLM proxy exposes more sycophantic collapse than pre-generated scripts. Among all tactics, emotional appeals is the most associated with inducing LLM sycophantic behavior. The code and data are released at https://anonymous.4open.science/r/SPINE

cs.CL↗

arXiv2026

TRACE: Trajectory Aware Reasoning for Multi-Turn Adversarial Conversation Evaluation

Multi-turn jailbreak attacks have emerged as a critical safety threat to LLMs, as harmful objectives are decomposed across a sequence of apparently benign turns to bypass guardrails. Existing defenses lack the reasoning capacity to identify evolving manipulation patterns, often trading helpfulness for safety by over-refusing benign requests related to sensitive topics. We introduce Trace, a multi-turn defense with trajectory-aware structured reasoning. Before generating each response, the model identifies manipulation cues from the trajectory, evaluates both the benign and adversarial interpretations of user intent, assigns a jailbreak score, and commits to an action: Allow, Caution, or Decline. We curate 4k multi-turn adversarial conversations from five attack frameworks, pair them with 2.4k benign dialogs, and 600 sensitive-but-benign conversations. We train Llama-3.1-8B-Instruct with SFT and GRPO under a multi-component reward that jointly optimizes helpfulness on benign prompts and robustness against jailbreak attempts. Across seven multi-turn attack benchmarks, Trace attains an average attack success rate (ASR) of 14.5% against 31.4% for the strongest baseline and 74.9% for the undefended target, while significantly raising the attacker effort required per successful jailbreak. Trace also balances usability and safety, achieving a 93.3% average compliance on over-refusal benchmarks.

cs.AI↗

arXiv2026

Sycophantic Praise: Evaluating Excessive Praise in Language Models

Sycophancy in language models is typically studied as excessive agreement or validation, while explicit praise and flattery have received comparatively little attention. We argue that sycophantic praise is a distinct alignment problem that cannot be reliably measured using current methods. We introduce a parameterized framework that measures whether praise is excessive relative to contribution quality and expected user ability. We show that our framework substantially outperforms generic LLM judges in agreement with human annotations, and that sycophantic praise occurs far more frequently in social and interpretive domains than in objective reasoning settings. Together, these findings position praise calibration as a distinct alignment challenge.

cs.CL↗

arXiv2026

MMR-GRPO: Accelerating GRPO-Style Training through Diversity-Aware Reward Reweighting

Group Relative Policy Optimization (GRPO) has become a standard approach for training mathematical reasoning models; however, its reliance on multiple completions per prompt makes training computationally expensive. Although recent work has reduced the number of training steps required to reach peak performance, the overall wall-clock training time often remains unchanged or even increases due to higher per-step cost. We propose MMR-GRPO, which integrates Maximal Marginal Relevance to reweigh rewards based on completion diversity. Our key insight is that semantically redundant completions contribute limited marginal learning signal; prioritizing diverse solutions yields more informative updates and accelerates convergence. Extensive evaluations across three model sizes (1.5B, 7B, 8B), three GRPO variants, and five mathematical reasoning benchmarks show that MMR-GRPO achieves comparable peak performance while requiring on average 47.9% fewer training steps and 70.2% less wall-clock time. These gains are consistent across models, methods, and benchmarks. Our code is released at: https://github.com/WeiKangda/MMR-GRPO.

cs.LG↗

arXiv2025

Thermal Analysis of 3D GPU-Memory Architectures with Boron Nitride Interposer

As artificial intelligence (AI) chips become more powerful, the thermal management capabilities of conventional silicon (Si) substrates become insufficient for 3D-stacked designs. This work integrates electrically insulative and thermally conductive hexagonal boron nitride (h-BN) interposers into AI chips for effective thermal management. Using COMSOL Multiphysics, the effects of High-Bandwidth Memory (HBM) distributions and thermal interface material configurations on heat dissipation and hotspot mitigation were studied. A 20 {\deg}C reduction in hot spots was achieved using h-BN interposers compared to Si interposers. Such an improvement could reduce AI chips' power leakage by 22% and significantly enhance their thermal performance.

eess.SP↗

arXiv2025

Multi-document Summarization through Multi-document Event Relation Graph Reasoning in LLMs: a case study in Framing Bias Mitigation

Media outlets are becoming more partisan and polarized nowadays. Most previous work focused on detecting media bias. In this paper, we aim to mitigate media bias by generating a neutralized summary given multiple articles presenting different ideological views. Motivated by the critical role of events and event relations in media bias detection, we propose to increase awareness of bias in LLMs via multi-document events reasoning and use a multi-document event relation graph to guide the summarization process. This graph contains rich event information useful to reveal bias: four common types of in-doc event relations to reflect content framing bias, cross-doc event coreference relation to reveal content selection bias, and event-level moral opinions to highlight opinionated framing bias. We further develop two strategies to incorporate the multi-document event relation graph for neutralized summarization. Firstly, we convert a graph into natural language descriptions and feed the textualized graph into LLMs as a part of a hard text prompt. Secondly, we encode the graph with graph attention network and insert the graph embedding into LLMs as a soft prompt. Both automatic evaluation and human evaluation confirm that our approach effectively mitigates both lexical and informational media bias, and meanwhile improves content preservation.

cs.CL↗

arXiv2025

Hidden in Plain Sight: Evaluation of the Deception Detection Capabilities of LLMs in Multimodal Settings

Detecting deception in an increasingly digital world is both a critical and challenging task. In this study, we present a comprehensive evaluation of the automated deception detection capabilities of Large Language Models (LLMs) and Large Multimodal Models (LMMs) across diverse domains. We assess the performance of both open-source and commercial LLMs on three distinct datasets: real life trial interviews (RLTD), instructed deception in interpersonal scenarios (MU3D), and deceptive reviews (OpSpam). We systematically analyze the effectiveness of different experimental setups for deception detection, including zero-shot and few-shot approaches with random or similarity-based in-context example selection. Our results show that fine-tuned LLMs achieve state-of-the-art performance on textual deception detection tasks, while LMMs struggle to fully leverage cross-modal cues. Additionally, we analyze the impact of auxiliary features, such as non-verbal gestures and video summaries, and examine the effectiveness of different prompting strategies, including direct label generation and chain-of-thought reasoning. Our findings provide key insights into how LLMs process and interpret deceptive cues across modalities, highlighting their potential and limitations in real-world deception detection applications.

cs.CL↗

arXiv2025

Mitigating Gender Bias via Fostering Exploratory Thinking in LLMs

Large Language Models (LLMs) often exhibit gender bias, resulting in unequal treatment of male and female subjects across different contexts. To address this issue, we propose a novel data generation framework that fosters exploratory thinking in LLMs. Our approach prompts models to generate story pairs featuring male and female protagonists in structurally identical, morally ambiguous scenarios, then elicits and compares their moral judgments. When inconsistencies arise, the model is guided to produce balanced, gender-neutral judgments. These story-judgment pairs are used to fine-tune or optimize the models via Direct Preference Optimization (DPO). Experimental results show that our method significantly reduces gender bias while preserving or even enhancing general model capabilities. We will release the code and generated data. We release the code and generated data at: https://github.com/WeiKangda/LLMs-Exploratory-Bias-Mitigation/tree/main.

cs.CL↗

arXiv2025

CliME: Evaluating Multimodal Climate Discourse on Social Media and the Climate Alignment Quotient (CAQ)

The rise of Large Language Models (LLMs) has raised questions about their ability to understand climate-related contexts. Though climate change dominates social media, analyzing its multimodal expressions is understudied, and current tools have failed to determine whether LLMs amplify credible solutions or spread unsubstantiated claims. To address this, we introduce CliME (Climate Change Multimodal Evaluation), a first-of-its-kind multimodal dataset, comprising 2579 Twitter and Reddit posts. The benchmark features a diverse collection of humorous memes and skeptical posts, capturing how these formats distill complex issues into viral narratives that shape public opinion and policy discussions. To systematically evaluate LLM performance, we present the Climate Alignment Quotient (CAQ), a novel metric comprising five distinct dimensions: Articulation, Evidence, Resonance, Transition, and Specificity. Additionally, we propose three analytical lenses: Actionability, Criticality, and Justice, to guide the assessment of LLM-generated climate discourse using CAQ. Our findings, based on the CAQ metric, indicate that while most evaluated LLMs perform relatively well in Criticality and Justice, they consistently underperform on the Actionability axis. Among the models evaluated, Claude 3.7 Sonnet achieves the highest overall performance. We publicly release our CliME dataset and code to foster further research in this domain.

cs.CL↗

arXiv2025

PreMind: Multi-Agent Video Understanding for Advanced Indexing of Presentation-style Videos

In recent years, online lecture videos have become an increasingly popular resource for acquiring new knowledge. Systems capable of effectively understanding/indexing lecture videos are thus highly desirable, enabling downstream tasks like question answering to help users efficiently locate specific information within videos. This work proposes PreMind, a novel multi-agent multimodal framework that leverages various large models for advanced understanding/indexing of presentation-style videos. PreMind first segments videos into slide-presentation segments using a Vision-Language Model (VLM) to enhance modern shot-detection techniques. Each segment is then analyzed to generate multimodal indexes through three key steps: (1) extracting slide visual content, (2) transcribing speech narratives, and (3) consolidating these visual and speech contents into an integrated understanding. Three innovative mechanisms are also proposed to improve performance: leveraging prior lecture knowledge to refine visual understanding, detecting/correcting speech transcription errors using a VLM, and utilizing a critic agent for dynamic iterative self-reflection in vision analysis. Compared to traditional video indexing methods, PreMind captures rich, reliable multimodal information, allowing users to search for details like abbreviations shown only on slides. Systematic evaluations on the public LPM dataset and an internal enterprise dataset are conducted to validate PreMind's effectiveness, supported by detailed analyses.

cs.CV↗

arXiv2025

LegalCore: A Dataset for Event Coreference Resolution in Legal Documents

Recognizing events and their coreferential mentions in a document is essential for understanding semantic meanings of text. The existing research on event coreference resolution is mostly limited to news articles. In this paper, we present the first dataset for the legal domain, LegalCore, which has been annotated with comprehensive event and event coreference information. The legal contract documents we annotated in this dataset are several times longer than news articles, with an average length of around 25k tokens per document. The annotations show that legal documents have dense event mentions and feature both short-distance and super long-distance coreference links between event mentions. We further benchmark mainstream Large Language Models (LLMs) on this dataset for both event detection and event coreference resolution tasks, and find that this dataset poses significant challenges for state-of-the-art open-source and proprietary LLMs, which perform significantly worse than a supervised baseline. We will publish the dataset as well as the code.

cs.CL↗

arXiv2024

Boosting Logical Fallacy Reasoning in LLMs via Logical Structure Tree

Logical fallacy uses invalid or faulty reasoning in the construction of a statement. Despite the prevalence and harmfulness of logical fallacies, detecting and classifying logical fallacies still remains a challenging task. We observe that logical fallacies often use connective words to indicate an intended logical relation between two arguments, while the argument semantics does not actually support the logical relation. Inspired by this observation, we propose to build a logical structure tree to explicitly represent and track the hierarchical logic flow among relation connectives and their arguments in a statement. Specifically, this logical structure tree is constructed in an unsupervised manner guided by the constituency tree and a taxonomy of connectives for ten common logical relations, with relation connectives as non-terminal nodes and textual arguments as terminal nodes, and the latter are mostly elementary discourse units. We further develop two strategies to incorporate the logical structure tree into LLMs for fallacy reasoning. Firstly, we transform the tree into natural language descriptions and feed the textualized tree into LLMs as a part of the hard text prompt. Secondly, we derive a relation-aware tree embedding and insert the tree embedding into LLMs as a soft prompt. Experiments on benchmark datasets demonstrate that our approach based on logical structure tree significantly improves precision and recall for both fallacy detection and fallacy classification.

cs.CL↗

arXiv2024

Evaluating Gender Bias of LLMs in Making Morality Judgements

Large Language Models (LLMs) have shown remarkable capabilities in a multitude of Natural Language Processing (NLP) tasks. However, these models are still not immune to limitations such as social biases, especially gender bias. This work investigates whether current closed and open-source LLMs possess gender bias, especially when asked to give moral opinions. To evaluate these models, we curate and introduce a new dataset GenMO (Gender-bias in Morality Opinions) comprising parallel short stories featuring male and female characters respectively. Specifically, we test models from the GPT family (GPT-3.5-turbo, GPT-3.5-turbo-instruct, GPT-4-turbo), Llama 3 and 3.1 families (8B/70B), Mistral-7B and Claude 3 families (Sonnet and Opus). Surprisingly, despite employing safety checks, all production-standard models we tested display significant gender bias with GPT-3.5-turbo giving biased opinions in 24% of the samples. Additionally, all models consistently favour female characters, with GPT showing bias in 68-85% of cases and Llama 3 in around 81-85% instances. Additionally, our study investigates the impact of model parameters on gender bias and explores real-world situations where LLMs reveal biases in moral decision-making.

cs.CL↗

arXiv2024

OD-Stega: LLM-Based Relatively Secure Steganography via Optimized Distributions

We consider coverless steganography where a Large Language Model (LLM) is used to generate stego-texts in combination with arithmetic coding. An efficient method should embed secret bits in as few language tokens as possible while keeping the stego-text as natural as possible. We show that this problem is equivalent to maximizing the entropy of a replacement probability distribution of the next token generation, subject to a constraint on the divergence between the new distribution and the original one produced by the LLM. A closed-form solution is provided under either the KL divergence or the total variation constraint. Several important practical issues are also tackled: 1) An often-overlooked tokenization mismatch issue is resolved with a simple prompt selection approach, 2) The combination of the optimized distribution and the vocabulary truncation technique is considered, and 3) The incorporation of the proposed approach with existing (potentially non arithmetic coding based) techniques, e.g., the Discop technique.

cs.IT↗

arXiv2024

UAL-Bench: The First Comprehensive Unusual Activity Localization Benchmark

Localizing unusual activities, such as human errors or surveillance incidents, in videos holds practical significance. However, current video understanding models struggle with localizing these unusual events likely because of their insufficient representation in models' pretraining datasets. To explore foundation models' capability in localizing unusual activity, we introduce UAL-Bench, a comprehensive benchmark for unusual activity localization, featuring three video datasets: UAG-OOPS, UAG-SSBD, UAG-FunQA, and an instruction-tune dataset: OOPS-UAG-Instruct, to improve model capabilities. UAL-Bench evaluates three approaches: Video-Language Models (Vid-LLMs), instruction-tuned Vid-LLMs, and a novel integration of Vision-Language Models and Large Language Models (VLM-LLM). Our results show the VLM-LLM approach excels in localizing short-span unusual events and predicting their onset (start time) more accurately than Vid-LLMs. We also propose a new metric, R@1, TD <= p, to address limitations in existing evaluation methods. Our findings highlight the challenges posed by long-duration videos, particularly in autism diagnosis scenarios, and the need for further advancements in localization techniques. Our work not only provides a benchmark for unusual activity localization but also outlines the key challenges for existing foundation models, suggesting future research directions on this important task.

cs.CV↗

arXiv2024

Are LLMs Good Annotators for Discourse-level Event Relation Extraction?

Large Language Models (LLMs) have demonstrated proficiency in a wide array of natural language processing tasks. However, its effectiveness over discourse-level event relation extraction (ERE) tasks remains unexplored. In this paper, we assess the effectiveness of LLMs in addressing discourse-level ERE tasks characterized by lengthy documents and intricate relations encompassing coreference, temporal, causal, and subevent types. Evaluation is conducted using an commercial model, GPT-3.5, and an open-source model, LLaMA-2. Our study reveals a notable underperformance of LLMs compared to the baseline established through supervised learning. Although Supervised Fine-Tuning (SFT) can improve LLMs performance, it does not scale well compared to the smaller supervised baseline model. Our quantitative and qualitative analysis shows that LLMs have several weaknesses when applied for extracting event relations, including a tendency to fabricate event mentions, and failures to capture transitivity rules among relations, detect long distance relations, or comprehend contexts with dense event mentions. Code available at: https://github.com/WeiKangda/LLM-ERE.git.

cs.CL↗

arXiv2024

LLMs Assist NLP Researchers: Critique Paper (Meta-)Reviewing

This work is motivated by two key trends. On one hand, large language models (LLMs) have shown remarkable versatility in various generative tasks such as writing, drawing, and question answering, significantly reducing the time required for many routine tasks. On the other hand, researchers, whose work is not only time-consuming but also highly expertise-demanding, face increasing challenges as they have to spend more time reading, writing, and reviewing papers. This raises the question: how can LLMs potentially assist researchers in alleviating their heavy workload? This study focuses on the topic of LLMs assist NLP Researchers, particularly examining the effectiveness of LLM in assisting paper (meta-)reviewing and its recognizability. To address this, we constructed the ReviewCritique dataset, which includes two types of information: (i) NLP papers (initial submissions rather than camera-ready) with both human-written and LLM-generated reviews, and (ii) each review comes with "deficiency" labels and corresponding explanations for individual segments, annotated by experts. Using ReviewCritique, this study explores two threads of research questions: (i) "LLMs as Reviewers", how do reviews generated by LLMs compare with those written by humans in terms of quality and distinguishability? (ii) "LLMs as Metareviewers", how effectively can LLMs identify potential issues, such as Deficient or unprofessional review segments, within individual paper reviews? To our knowledge, this is the first work to provide such a comprehensive analysis.

cs.CL↗

arXiv2024

Indexing Analytics to Instances: How Integrating a Dashboard can Support Design Education

We investigate how to use AI-based analytics to support design education. The analytics at hand measure multiscale design, that is, students' use of space and scale to visually and conceptually organize their design work. With the goal of making the analytics intelligible to instructors, we developed a research artifact integrating a design analytics dashboard with design instances, and the design environment that students use to create them. We theorize about how Suchman's notion of mutual intelligibility requires contextualized investigation of AI in order to develop findings about how analytics work for people. We studied the research artifact in 5 situated course contexts, in 3 departments. A total of 236 students used the multiscale design environment. The 9 instructors who taught those students experienced the analytics via the new research artifact. We derive findings from a qualitative analysis of interviews with instructors regarding their experiences. Instructors reflected on how the analytics and their presentation in the dashboard have the potential to affect design education. We develop research implications addressing: (1) how indexing design analytics in the dashboard to actual design work instances helps design instructors reflect on what they mean and, more broadly, is a technique for how AI-based design analytics can support instructors' assessment and feedback experiences in situated course contexts; and (2) how multiscale design analytics, in particular, have the potential to support design education. By indexing, we mean linking which provides context, here connecting the numbers of the analytics with visually annotated design work instances.

cs.HC↗