A core part of scientific peer review involves providing expert critiques that directly assess the scientific claims a paper makes. While it is now possible to automatically generate plausible (if generic) reviews, ensuring that these reviews are sound and grounded in the papers’ claims remains challenging. To facilitate LLM benchmarking on these challenges, we introduce CLAIMCHECK, an annotated dataset of NeurIPS 2023 and 2024 submissions and reviews mined from OpenReview. CLAIMCHECK is richly annotated by ML experts for weakness statements in the reviews and the paper claims that they dispute, as well as fine-grained labels of the validity, objectivity, and type of the identified weaknesses. We benchmark several LLMs on three claim-centric tasks supported by CLAIMCHECK, requiring models to (1) associate weaknesses with the claims they dispute, (2) predict fine-grained labels for weaknesses and rewrite the weaknesses to enhance their specificity, and (3) verify a paper’s claims with grounded reasoning. Our experiments reveal that cutting-edge LLMs, while capable of predicting weakness labels in (2), continue to underperform relative to human experts on all other tasks.
论文检索
输入标题、作者或关键词，从 7,591 篇学术成果中精准定位
Research in uncertainty quantification (UQ) for large language models (LLMs) is increasingly important towards guaranteeing the reliability of this groundbreaking technology. We explore the integration of LLM UQ methods in argumentative LLMs (ArgLLMs), an explainable LLM framework for decision-making based on computational argumentation in which UQ plays a critical role. We conduct experiments to evaluate ArgLLMs’ performance on claim verification tasks when using different LLM UQ methods, inherently performing an assessment of the UQ methods’ effectiveness. Moreover, the experimental procedure itself is a novel way of evaluating the effectiveness of UQ methods, especially when intricate and potentially contentious statements are present. Our results demonstrate that, despite its simplicity, direct prompting is an effective UQ strategy in ArgLLMs, outperforming considerably more complex approaches.
TriSPrompt: A Hierarchical Soft Prompt Model for Multimodal Rumor Detection with Incomplete Modalities
PDF ↗The widespread presence of incomplete modalities in multimodal data poses a significant challenge to achieving accurate rumor detection. Existing multimodal rumor detection methods primarily focus on learning joint modality representations from complete multimodal training data, rendering them ineffective in addressing the common occurrence of missing modalities in real-world scenarios. In this paper, we propose a hierarchical soft prompt model TriSPrompt, which integrates three types of prompts, i.e., modality-aware (MA) prompt, modality-missing (MM) prompt, and mutual-views (MV) prompt, to effectively detect rumors in incomplete multimodal data. The MA prompt captures both heterogeneous information from specific modalities and homogeneous features from available data, aiding in modality recovery. The MM prompt models missing states in incomplete data, enhancing the model’s adaptability to missing information. The MV prompt learns relationships between subjective (i.e., text and image) and objective (i.e., comments) perspectives, effectively detecting rumors. Extensive experiments on three real-world benchmarks demonstrate that TriSPrompt achieves an accuracy gain of over 13% compared to state-of-the-art methods. The codes and datasets are available at https: //anonymous.4open.science/r/code-3E88.
Transformers have achieved great success in effectively processing sequential data such as text. Their architecture consisting of several attention and feedforward blocks can model relations between elements of a sequence in parallel manner, which makes them very efficient to train and effective in sequence modeling. Even though they have shown strong performance in processing sequential data, the size of their parameters is considerably larger when compared to other architectures such as RNN and CNN based models. Therefore, several approaches have explored parameter sharing and recurrence in Transformer models to address their computational demands. However, such methods struggle to maintain high performance compared to the original transformer model. To address this challenge, we propose our novel approach, RingFormer, which employs one Transformer layer that processes input repeatedly in a circular, ring-like manner, while utilizing low-rank matrices to generate input-dependent level signals. This allows us to reduce the model parameters substantially while maintaining high performance in a variety of tasks such as translation and image classification, as validated in the experiments.
PVTNL: Prompting Vision Transformers with Natural Language for Generalizable Person Re-identification
PDF ↗Domain generalization person re-identification (DG-ReID) aims to train models on source domains and generalize to unseen target domains.While patch-based Vision Transformers have achieved success in capturing fine-grained visual features, they often overlook global semantic structure and suffer from feature entanglement, leading to overfitting across domains. Meanwhile, natural language provides high-level semantic abstraction but lacks spatial precision for fine-grained alignment.We propose PVTNL (Prompting Vision Transformers with Natural Language), a novel framework for generalizable person re-identification. PVTNL leverages the pre-trained vision-language model BLIP to extract aligned visual and textual embeddings. Specifically, we utilize body-part cues to segment images into semantically coherent regions and align them with corresponding natural language descriptions. These region-level textual prompts are encoded and injected as soft prompts into the Vision Transformer to guide localized feature learning. Notably, our language module is retained during inference, enabling persistent semantic grounding that enhances cross-domain generalization.Extensive experiments on standard DG-ReID benchmarks demonstrate that PVTNL achieves state-of-the-art performance. Ablation studies further confirm the effectiveness of body-part-level alignment, soft language prompting, and the benefit of preserving language guidance at inference time.
Object detection is a core challenge in computer vision. Traditional methods primarily rely on intermediate modalities such as text, speech, or visual cues to interpret user intent, leading to inefficient and potentially distorted expressions of intent. Brain signals, particularly fMRI signals, emerge as a novel modality that can directly reflect user intent, eliminating ambiguities introduced during modality conversion. However, brain signal-based object detection still faces challenges in accuracy and robustness. To address these challenges, we present BrainLoc, a lightweight object detection model guided by fMRI signals. First, we employ a multi-modal alignment strategy that enhances fMRI signal feature extraction by incorporating various modalities including images and text. Second, we propose a cross-domain fusion module that promotes interaction between fMRI features and category features, improving the representation of category information in fMRI signals. Extensive experiments demonstrate that BrainLoc achieves state-of-the-art performance in brain signal-based object detection tasks, showing significant advantages in both accuracy and convenience.
Planning in modern LLM agents relies on the utilization of LLM as an internal world model, acquired during pretraining. However, existing agent designs fail to effectively assimilate new observations into dynamic updates of the world model. This reliance on the LLM’s static internal world model is progressively prone to misalignment with the underlying true state of the world, leading to the generation of divergent and erroneous plans. We introduce a hierarchical agent architecture, CoEx, in which hierarchical state abstraction allows LLM planning to co-evolve with a dynamically updated model of the world. CoEx plans and interacts with the world by using LLM reasoning to orchestrate dynamic plans consisting of subgoals, and its learning mechanism continuously incorporates these subgoal experiences into a persistent world model in the form of a neurosymbolic belief state, comprising textual inferences and code-based symbolic memory. We evaluate our agent across a diverse set of agent scenarios involving rich environments and complex tasks including ALFWorld, PDDL, and Jericho. Our experiments show that CoEx outperforms existing agent paradigms in planning and exploration.
Do LLMs Align Human Values Regarding Social Biases? Judging and Explaining Social Biases with LLMs
PDF ↗Large language models (LLMs) can lead to undesired consequences when misaligned with human values, especially in scenarios involving complex and sensitive social biases. Previous studies have revealed the misalignment of LLMs with human values using expert-designed or agent-based emulated bias scenarios. However, it remains unclear whether the alignment of LLMs with human values differs across different types of scenarios (e.g., scenarios containing negative vs. non-negative questions). In this study, we investigate the alignment of LLMs with human values regarding social biases (HVSB) in different types of bias scenarios. Through extensive analysis of 12 LLMs from four model families and four datasets, we demonstrate that LLMs with large model parameter scales do not necessarily have lower misalignment rate and attack success rate. Moreover, LLMs show a certain degree of alignment preference for specific types of scenarios and the LLMs from the same model family tend to have higher judgment consistency. In addition, we study the understanding capacity of LLMs with their explanations of HVSB. We find no significant differences in the understanding of HVSB across LLMs. We also find LLMs prefer their own generated explanations. Additionally, we endow smaller language models (LMs) with the ability to explain HVSB.The generation results show that the explanations generated by the fine-tuned smaller LMs are more readable, but have a relatively lower agreeability.
Maltese is a unique Semitic language that has evolved under extensive influence from Romance and Germanic languages, particularly Italian and English. Despite its Semitic roots, its orthography is based on the Latin script, creating a gap between it and its closest linguistic relatives in Arabic. In this paper, we explore whether Arabic-language resources can support Maltese natural language processing (NLP) through cross-lingual augmentation techniques. We investigate multiple strategies for aligning Arabic textual data with Maltese, including various transliteration schemes and machine translation (MT) approaches. As part of this, we also introduce novel transliteration systems that better represent Maltese orthography. We evaluate the impact of these augmentations on monolingual and mutlilingual models and demonstrate that Arabic-based augmentation can significantly benefit Maltese NLP tasks.
Large Multimodal Models (LMMs) are increasingly vulnerable to AI-generated extremist content, including photorealistic images and text, which can be used to bypass safety mechanisms and generate harmful outputs. However, existing datasets for evaluating LMM robustness offer limited exploration of extremist content, often lacking AI-generated images, diverse image generation models, and comprehensive coverage of historical events, which hinders a complete assessment of model vulnerabilities. To fill this gap, we introduce ExtremeAIGC, a benchmark dataset and evaluation framework designed to assess LMM vulnerabilities against such content. ExtremeAIGC simulates real-world events and malicious use cases by curating diverse text and image based examples crafted using state-of-the-art image generation techniques. Our study reveals alarming weaknesses in LMMs, demonstrating that even cutting-edge safety measures fail to prevent the generation of extremist material. We systematically quantify the success rates of various attack strategies, exposing critical gaps in current defenses and emphasizing the need for more robust mitigation strategies. The code and data can be found at https://github.com/TheProParadox/ExtremeAIGC.
Large language models (LLMs) make it easy to rewrite a text in any style – e.g. to make it more polite, persuasive, or more positive – but evaluation thereof is not straightforward. A challenge lies in measuring content preservation: that content not attributable to style change is retained. This paper presents a large meta-evaluation of metrics for evaluating style and attribute transfer, focusing on content preservation. We find that meta-evaluation studies on existing datasets lead to misleading conclusions about the suitability of metrics for content preservation. Widely used metrics show a high correlation with human judgments despite being deemed unsuitable for the task – because they do not abstract from style changes when evaluating content preservation. We show that the overly high correlations with human judgment stem from the nature of the test data. To address this issue, we introduce a new, challenging test set specifically designed for evaluating content preservation metrics for style transfer. We construct the data by creating high variation in the content preservation. Using this dataset, we demonstrate that suitable metrics for content preservation for style transfer indeed are style-aware.To support efficient evaluation, we propose a new style-aware method that utilises small language models, obtaining a higher alignment with human judgements than prompting a model of a similar size as an autorater.
Deciphering molecular meaning in chemistry and biomedicine depends on context — a capability that large language models (LLMs) can enhance by aligning molecular structures with language. However, existing molecule-text models ignore complementary information in different molecular views and rely on single-view representations, limiting molecule structural understanding. Moreover, naïve multi-view alignment strategies face two challenges: (1) the aligned spaces differ across views due to inconsistent molecule-text mappings, and (2) existing loss objectives fail to preserve complementary information necessary for finegrained alignment. To enhance LLM’s ability to understand molecular structure, we propose MV-CLAM, a novel framework that aligns multi-view molecular representations into a unified textual space using a multi-querying transformer (MQ-Former). Our approach ensures cross-view consistency while the proposed token-level contrastive loss preserves diverse molecular features across textual queries. MV-CLAM enhances molecular reasoning, improving retrieval and captioning accuracy. The source code of MV-CLAM is available in https://github.com/sumin124/mv-clam.
Self-disclosure can provide psychological comfort and social support, but it also carries the risk of unintentionally revealing sensitive information, leading to serious privacy concerns. Research on self-disclosure in Chinese multimodal contexts remains limited, lacking high-quality corpora, analysis, and methods for detection. This work focuses on self-disclosure behaviors on Chinese multimodal social media platforms and constructs a high-quality text-image corpus to address this critical data gap. We systematically analyze the distribution of self-disclosure types, modality preferences, and their relationship with user intent, uncovering expressive patterns unique to the Chinese multimodal context. We also fine-tune five multimodal large language models to enhance self-disclosure detection in multimodal scenarios. Among these models, the Qwen2.5-omni-7B achieved a strong performance, with a partial span F1 score of 88.2%. This study provides a novel research perspective on multimodal self-disclosure in the Chinese context.
Composing language models (LMs) into multi-step language programs and automatically optimizing their modular prompts is now a mainstream paradigm for building AI systems, but the tradeoffs in this space have only scarcely been studied before. We introduce LangProBe, the first large-scale benchmark for evaluating the architectures and optimization strategies for language programs, with over 2000 combinations of tasks, architectures, optimizers, and choices of LMs. Using LangProBe, we are the first to study the impact of program architectures and optimizers (and their compositions together and with different models) on tradeoffs of quality and cost. We find that optimized language programs offer strong cost-quality Pareto improvement over raw calls to models, but simultaneously demonstrate that human judgment (or empirical decisions) about which compositions to pursue is still necessary for best performance.
Creative writing is a key capability of Large Language Models (LLMs), with potential applications in literature, storytelling, and various creative domains. However, evaluating the creativity of machine-generated texts remains a significant challenge, as existing methods either rely on costly manual annotations or fail to align closely with human assessments. In this paper, we propose an effective automated evaluation method based on the Torrance Test of Creative Writing (TTCW), which evaluates creativity as product. Our method employs a reference-based Likert-style approach, scoring generated creative texts relative to high-quality reference texts across various tests. Experimental results demonstrate that our method significantly improves the alignment between LLM evaluations and human assessments, achieving a pairwise accuracy of 0.75 (+15%).
Automated Claim Verification (CV)—the task of assessing a claim’s veracity against explicitly provided evidence—is a critical tool in the fight against growing misinformation. This survey offers a comprehensive analysis of 198 studies published between January 2022 and March 2025, synthesizing recent advances in CV corpus creation and system design. Through two in-depth case studies, we illuminate persistent challenges in veracity annotation, limitations of conventional CV pipelines, and pitfalls in recent claim decomposition approaches. We conclude by identifying key unresolved challenges and proposing productive directions for future research.
Recent large language models have shown promising capabilities in long-form reasoning, following structured chains of thought before arriving at a final answer. However, we observe that these reasoning paths tend to include substantial redundancy; analyzing attention patterns reveals that attention scores are widely scattered, particularly incorrect answers exhibit greater attention sparsity. In this paper, we demonstrate that deliberately removing this redundancy in the reasoning process significantly improves the performance through clear thinking (i.e., removing distraction). Specifically, we systematically identify such redundancy by measuring token-level attention scores to a special end-of-thinking token, which is appended to an explicit instruction inserted to conclude each intermediate reasoning step. Furthermore, we propose structure-aware pruning that prioritizes removing tokens in low-contributing reasoning chunks over individual tokens. After evicting redundant tokens, we remove the injected end-of-thinking instruction, then resume the reasoning generation. We demonstrate that our method significantly improves the over all accuracy across reasoning-intensive benchmarks without any training involved. In particular, our method shows strong performance on challenging mathematics competition benchmarks such as AIME and AMC, where reasoning redundancy is more prevalent.
Expanding the long-context capabilities of Multi-modal Large Language Models (MLLMs) is critical for advancing video understanding and high-resolution image analysis. Achieving this requires systematic improvements in model architecture, data construction, and training strategies, particularly to address challenges such as performance degradation with increasing image counts and high computational costs. In this paper, we propose a hybrid architecture that integrates Mamba and Transformer blocks, introduce data construction methods that capture both temporal and spatial dependencies, and employ a progressive training strategy. Our released model, LongLLaVA (Long-Context Large Language and Vision Assistant), demonstrates an effective balance between efficiency and performance. LongLLaVA achieves competitive results across various benchmarks while maintaining high throughput and low memory consumption. Notably, it can process nearly one thousand images on a single A100 80GB GPU, underscoring its potential for a wide range of multi-modal applications.
Speculative decoding has emerged as a promising approach to accelerating large language model (LLM) generation using a fast drafter while maintaining alignment with the target model’s distribution. However, existing approaches face a trade-off: external drafters offer flexibility but can suffer from slower drafting, while self-speculation methods use drafters tailored to the target model but require re-training. In this paper, we introduce novel drafters based on Mamba, a state-of-the-art state space model (SSM), as a solution that combines the best aspects of both approaches. By leveraging the linear structure of SSMs, our approach avoids the quadratic complexity inherent in traditional Transformer-based methods, enabling faster drafting and lower memory usage while maintaining the flexibility to work across different target models. We further enhance efficiency with a novel test-time tree search algorithm for generating high-quality draft candidates. Our empirical evaluation demonstrates that Mamba-based drafters not only outperform existing external drafting methods but are also comparable to state-of-the-art self-speculation approaches while using less memory and maintaining their cross-model adaptability.
Text style transfer (TST) modifies a source sentence to match a target style while preserving its semantics. While existing models perform well on simple styles like sentiment and formality, they struggle with complex, entangled styles such as poetry and brand-specific tones, which require advanced operations to disentangle content and style. We propose a multi-agent self-check framework that contains a large language model (LLM) as a planner for disentangling subtasks and expert agents for executing the subtasks. This training-free multi-agent framework decomposes TST into manageable components, enabling iterative refinement through a self-check module that balances style adherence and content preservation. Experiments on both simple and complex style datasets show our framework significantly improves style strength and content preservation, with strong adaptability in few-shot settings.