Zhecheng Li

Zhecheng Li, Yiwei Wang, Bryan Hooi et al.

Question answering represents a core capability of large language models (LLMs). However, when individuals encounter unfamiliar knowledge in texts, they often formulate questions that the text itself cannot answer due to insufficient understanding of the underlying information. Recent studies reveal that while LLMs can detect unanswerable questions, they struggle to assist users in reformulating these questions. Even advanced models like GPT-3.5 demonstrate limited effectiveness in this regard. To address this limitation, we propose DRS: Deep Question Reformulation with Structured Output, a novel zero-shot method aimed at enhancing LLMs ability to assist users in reformulating questions to extract relevant information from new documents. DRS combines the strengths of LLMs with a DFS-based algorithm to iteratively explore potential entity combinations and constrain outputs using predefined entities. This structured approach significantly enhances the reformulation capabilities of LLMs. Comprehensive experimental evaluations demonstrate that DRS improves the reformulation accuracy of GPT-3.5 from $23.03\%$ to $70.42\%$, while also enhancing the performance of open-source models, such as Gemma2-9B, from $26.35\%$ to $56.75\%$.

Zhen Xiong, Yujun Cai, Bryan Hooi et al.

Large Language Models (LLMs) have demonstrated strong generalization capabilities across a wide range of natural language processing (NLP) tasks. However, they exhibit notable weaknesses in character-level string manipulation, struggling with fundamental operations such as character deletion, insertion, and substitution. These challenges stem primarily from tokenization constraints, despite the critical role of such operations in data preprocessing and code generation. Through systematic analysis, we derive two key insights: (1) LLMs face significant difficulties in leveraging intrinsic token knowledge for character-level reasoning, and (2) atomized word structures can substantially enhance LLMs' ability to process token-level structural information. Building on these insights, we propose Character-Level Manipulation via Divide and Conquer, a novel approach designed to bridge the gap between token-level processing and character-level manipulation. Our method decomposes complex operations into explicit character-level subtasks coupled with controlled token reconstruction phases, leading to significant improvements in accuracy. Without additional training, our method significantly improves accuracies on the $\texttt{Deletion}$, $\texttt{Insertion}$, and $\texttt{Substitution}$ tasks. To support further research, we open-source our implementation and benchmarks.

Kaituo Zhang, Zhen Xiong, Mingyu Zhong et al.

Tool-augmented reasoning has become a popular direction for LLM-based agents, and it is widely assumed to improve reasoning and reliability. However, we demonstrate that this consensus does not always hold: in the presence of semantic distractors, tool-augmented reasoning does not necessarily outperform native CoT. To explain this performance gap, we propose a Factorized Intervention Framework that isolates the cost of prompt formatting, the overhead of the tool-calling protocol, and the actual gain from executing tools. Our analysis reveals a critical tradeoff: under semantic noise, the gains from tools often fail to offset the "tool-use tax", which is the performance degradation introduced by the tool-calling protocol itself. To address this, we introduce G-STEP, a lightweight inference-time gate to mitigate protocol-induced errors. While this yields partial recovery, our findings suggest that more substantial improvements still require strengthening the model's intrinsic reasoning and tool-interaction capabilities.

Zhecheng Li, Guoxian Song, Yujun Cai et al.

Modern Vision-Language Models (VLMs) exhibit remarkable visual and linguistic capabilities, achieving impressive performance in various tasks such as image recognition and object localization. However, their effectiveness in fine-grained tasks remains an open question. In everyday scenarios, individuals encountering design materials, such as magazines, typography tutorials, research papers, or branding content, may wish to identify aesthetically pleasing fonts used in the text. Given their multimodal capabilities and free accessibility, many VLMs are often considered potential tools for font recognition. This raises a fundamental question: Do VLMs truly possess the capability to recognize fonts? To investigate this, we introduce the Font Recognition Benchmark (FRB), a compact and well-structured dataset comprising 15 commonly used fonts. FRB includes two versions: (i) an easy version, where 10 sentences are rendered in different fonts, and (ii) a hard version, where each text sample consists of the names of the 15 fonts themselves, introducing a stroop effect that challenges model perception. Through extensive evaluation of various VLMs on font recognition tasks, we arrive at the following key findings: (i) Current VLMs exhibit limited font recognition capabilities, with many state-of-the-art models failing to achieve satisfactory performance and being easily affected by the stroop effect introduced by textual information. (ii) Few-shot learning and Chain-of-Thought (CoT) prompting provide minimal benefits in improving font recognition accuracy across different VLMs. (iii) Attention analysis sheds light on the inherent limitations of VLMs in capturing semantic features.

Zhen Xiong, Yujun Cai, Zhecheng Li et al.

Recent advances in test-time scaling have enabled Large Language Models (LLMs) to display sophisticated reasoning abilities via extended Chain-of-Thought (CoT) generation. Despite their potential, these Reasoning LLMs (RLMs) often demonstrate counterintuitive and unstable behaviors, such as performance degradation under few-shot prompting, that challenge our current understanding of RLMs. In this work, we introduce a unified graph-based analytical framework for better modeling the reasoning processes of RLMs. Our method first clusters long, verbose CoT outputs into semantically coherent reasoning steps, then constructs directed reasoning graphs to capture contextual and logical dependencies among these steps. Through comprehensive analysis across models and prompting regimes, we reveal that structural properties, such as exploration density, branching, and convergence ratios, strongly correlate with reasoning accuracy. Our findings demonstrate how prompting strategies substantially reshape the internal reasoning structure of RLMs, directly affecting task outcomes. The proposed framework not only enables quantitative evaluation of reasoning quality beyond conventional metrics but also provides practical insights for prompt engineering and the cognitive analysis of LLMs. Code and resources will be released to facilitate future research in this direction.

Zhecheng Li, Yiwei Wang, Bryan Hooi et al.

Vertical text input is commonly encountered in various real-world applications, such as mathematical computations and word-based Sudoku puzzles. While current large language models (LLMs) have excelled in natural language tasks, they remain vulnerable to variations in text formatting. Recent research demonstrates that modifying input formats, such as vertically aligning words for encoder-based models, can substantially lower accuracy in text classification tasks. While easily understood by humans, these inputs can significantly mislead models, posing a potential risk of bypassing detection in real-world scenarios involving harmful or sensitive information. With the expanding application of LLMs, a crucial question arises: Do decoder-based LLMs exhibit similar vulnerabilities to vertically formatted text input? In this paper, we investigate the impact of vertical text input on the performance of various LLMs across multiple text classification datasets and analyze the underlying causes. Our findings are as follows: (i) Vertical text input significantly degrades the accuracy of LLMs in text classification tasks. (ii) Chain-of-Thought (CoT) reasoning does not help LLMs recognize vertical input or mitigate its vulnerability, but few-shot learning with careful analysis does. (iii) We explore the underlying cause of the vulnerability by analyzing the inherent issues in tokenization and attention matrices.

Jiayu Yao, Shenghua Liu, Yiwei Wang et al.

Multimodal Retrieval-Augmented Generation (RAG) systems have become essential in knowledge-intensive and open-domain tasks. As retrieval complexity increases, ensuring the robustness of these systems is critical. However, current RAG models are highly sensitive to the order in which evidence is presented, often resulting in unstable performance and biased reasoning, particularly as the number of retrieved items or modality diversity grows. This raises a central question: How does the position of retrieved evidence affect multimodal RAG performance? To answer this, we present the first comprehensive study of position bias in multimodal RAG systems. Through controlled experiments across text-only, image-only, and mixed-modality tasks, we observe a consistent U-shaped accuracy curve with respect to evidence position. To quantify this bias, we introduce the Position Sensitivity Index ($PSI_p$) and develop a visualization framework to trace attention allocation patterns across decoder layers. Our results reveal that multimodal interactions intensify position bias compared to unimodal settings, and that this bias increases logarithmically with retrieval range. These findings offer both theoretical and empirical foundations for position-aware analysis in RAG, highlighting the need for evidence reordering or debiasing strategies to build more reliable and equitable generation systems.

Zhen Xiong, Yujun Cai, Zhecheng Li et al.

Controllable generation is a fundamental task in NLP with many applications, providing a basis for function calling to agentic communication. However, even state-of-the-art autoregressive Large Language Models (LLMs) today exhibit unreliability when required to generate structured output. Inspired by the current new diffusion-based large language models (dLLM), we realize that the architectural difference, especially the global information-sharing mechanism for language modeling, may be the key to unlock next-level controllable generation. To explore the possibility, we propose Self-adaptive Schema Scaffolding ($S^3$), a novel framework that enables dLLM to stably generate reliable structured outputs (e.g., JSON) by utilizing its innate reverse reasoning capability and global context awareness. $S^3$ initiates a schematic template directly in the output context as a starting state for dLLM, offering a more robust and general method than intricate prompt optimization. Experiments demonstrate that our method substantially unlocks the dLLM's potential in controllable generation in terms of structure adherence, content fidelity, and faithfulness. These results establish new perspectives and practical pathways for deploying language models in controllable generation tasks.

Zhecheng Li, Yiwei Wang, Bryan Hooi et al.

Cross-lingual summarization (CLS) aims to generate a summary for the source text in a different target language. Currently, instruction-tuned large language models (LLMs) excel at various English tasks. However, unlike languages such as English, Chinese or Spanish, for those relatively low-resource languages with limited usage or data, recent studies have shown that LLMs' performance on CLS tasks remains unsatisfactory even with few-shot settings. This raises the question: Are LLMs capable of handling cross-lingual summarization tasks for low-resource languages? To resolve this question, we fully explore the potential of large language models on cross-lingual summarization task for low-resource languages through our four-step zero-shot method: Summarization, Improvement, Translation and Refinement (SITR) with correspondingly designed prompts. We test our proposed method with multiple LLMs on two well-known cross-lingual summarization datasets with various low-resource target languages. The results show that: i) GPT-3.5 and GPT-4 significantly and consistently outperform other baselines when using our zero-shot SITR methods. ii) By employing our proposed method, we unlock the potential of LLMs, enabling them to effectively handle cross-lingual summarization tasks for relatively low-resource languages.

Zhecheng Li, Guoxian Song, Yiwei Wang et al.

Grounding natural language queries in graphical user interfaces (GUIs) presents a challenging task that requires models to comprehend diverse UI elements across various applications and systems, while also accurately predicting the spatial coordinates for the intended operation. To tackle this problem, we propose GMS: Generalist Scanner Meets Specialist Locator, a synergistic coarse-to-fine framework that effectively improves GUI grounding performance. GMS leverages the complementary strengths of general vision-language models (VLMs) and small, task-specific GUI grounding models by assigning them distinct roles within the framework. Specifically, the general VLM acts as a 'Scanner' to identify potential regions of interest, while the fine-tuned grounding model serves as a 'Locator' that outputs precise coordinates within these regions. This design is inspired by how humans perform GUI grounding, where the eyes scan the interface and the brain focuses on interpretation and localization. Our whole framework consists of five stages and incorporates hierarchical search with cross-modal communication to achieve promising prediction results. Experimental results on the ScreenSpot-Pro dataset show that while the 'Scanner' and 'Locator' models achieve only $2.0\%$ and $3.7\%$ accuracy respectively when used independently, their integration within GMS framework yields an overall accuracy of $35.7\%$, representing a $10 \times$ improvement. Additionally, GMS significantly outperforms other strong baselines under various settings, demonstrating its robustness and potential for general-purpose GUI grounding.

Zhen Xiong, Yujun Cai, Zhecheng Li et al.

Recent Large Audio-Language Models (LALMs) have shown strong performance on various audio understanding tasks such as speech translation and Audio Q\&A. However, they exhibit significant limitations on challenging audio reasoning tasks in complex acoustic scenarios. These situations would greatly benefit from the use of acoustic tools like noise suppression, source separation, and precise temporal alignment, but current LALMs lack access to such tools. To address this limitation, we introduce Thinking-with-Sound (TwS), a framework that equips LALMs with Audio CoT by combining linguistic reasoning with on-the-fly audio-domain analysis. Unlike existing approaches that treat audio as static input, TwS enables models to actively think with audio signals, performing numerical analysis and digital manipulation through multimodal reasoning. To evaluate this approach, we construct MELD-Hard1k, a new robustness benchmark created by introducing various acoustic perturbations. Experiments reveal that state-of-the-art LALMs suffer dramatic performance degradation on MELD-Hard1k, with accuracy dropping by more than $50\%$ compared to clean audio. TwS achieves substantial improvements in robustness, demonstrating both effectiveness and scalability: small models gain $24.73\%$ absolute accuracy, with improvements scaling consistently up to $36.61\%$ for larger models. Our findings demonstrate that Audio CoT can significantly enhance robustness without retraining, opening new directions for developing more robust audio understanding systems.

Zhecheng Li, Guoxian Song, Yiwei Wang et al.

Img2LaTeX is a practically important task that involves translating mathematical expressions and structured visual content from images into LaTeX code. In recent years, vision-language models (VLMs) have achieved remarkable progress across a range of visual understanding tasks, largely due to their strong generalization capabilities. However, despite initial efforts to apply VLMs to the Img2LaTeX task, their performance remains suboptimal. Empirical evidence shows that VLMs can be challenged by fine-grained visual elements, such as subscripts and superscripts in mathematical expressions, which results in inaccurate LaTeX generation. To address this challenge, we propose $A^2R^2$: Advancing Img2LaTeX Conversion via Visual Reasoning with Attention-Guided Refinement, a framework that effectively integrates attention localization and iterative refinement within a visual reasoning framework, enabling VLMs to perform self-correction and progressively improve LaTeX generation quality. For effective evaluation, we introduce a new dataset, Img2LaTex-Hard-1K, consisting of 1,100 carefully curated and challenging examples designed to rigorously evaluate the capabilities of VLMs within this task domain. Extensive experimental results demonstrate that: (1) $A^2R^2$ significantly improves model performance across various evaluation metrics spanning both textual and visual levels; (2) Increasing the number of inference rounds yields notable performance gains, underscoring the potential of $A^2R^2$ in test-time scaling scenarios; (3) Ablation studies and further evaluations confirm the effectiveness of our approach and the synergy of its core components during inference.