In dialogue intent detection, the challenge of acquiring sufficient corpora and the high cost of manual annotation often lead to incorrectly labeled or unrepresentative samples, which can hinder the generalization ability of classification models. Additionally, as using large language models for generating synthetic samples for data augmentation becomes more common, these synthetic samples may exacerbate the problem by introducing additional noise due to the models’ limited prior knowledge. To address this challenge, this paper proposes an interpretable Sample Filter by Topic Modeling (SFTM) framework. By evaluating the diversity and authenticity of the samples, SFTM effectively reduces the quantity of real and synthetic samples while improving the performance of the classification models. Our codes are publicly available at https://github.com/gumbouh/SFTM.
论文检索
输入标题、作者或关键词，从 7,537 篇学术成果中精准定位
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.
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.
Beyond Function-Level Search: Repository-Aware Dual-Encoder Code Retrieval with Adversarial Verification
PDF ↗The escalating complexity of modern codebases has intensified the need for code retrieval systems capable of interpreting cross-component change intents—a capability fundamentally absent in conventional function-level search paradigms. While recent research has improved alignment between queries and code snippets, retrieving contextually relevant code for certain change request remains underexplored. To bridge this gap, we present RepoAlignBench, the first benchmark designed to evaluate repository-level code retrieval for change request-driven scenarios, encompassing 52k columns. The benchmark shifts the paradigm from function-centric retrieval to holistic repository analysis. In addition, we propose ReflectCode, an adversarial reflection-augmented dual-tower architecture featuring disentangled code_encoder and doc_encoder towers. Our framework dynamically integrates syntactic patterns, function dependency, and semantic expansion intent through LLM. Comprehensive evaluations demonstrate that ReflectCode achieves 12.2% Top-5 Accuracy and 7.1% Recall improvements over state-of-the-art baselines.
Compositional generalization benchmarks seek to assess whether learning agents can successfully combine familiar concepts in novel ways. COGS (Kim & Linzen 2020, COGS, EMNLP) provides a suite of such tasks in the area of interpretive semantics (mapping sentences to logical forms). A noteworthy finding for COGS is that model performance varies widely across tasks. In this paper, we argue that these performance differences reflect deep properties of these tasks. We focus on two COGS tasks: an easy task (models are generally successful) and a hard task (no present-day models get any traction). Using both experiments and conceptual analysis, we argue that the easy task requires only a single distributional generalization that is well-supported by the training data, whereas the hard task involves a learning target that is ambiguous or even contradicted by the training data. We additionally argue that pretraining can disambiguate the hard task without compromising the goal of testing compositional generalization. Overall, our findings offer practical guidance to designers of compositional generalization benchmarks and also yield new insights into the nature of compositionality itself.
Characterizing Positional Bias in Large Language Models: A Multi-Model Evaluation of Prompt Order Effects
PDF ↗Large Language Models (LLMs) are widely used for a variety of tasks such as text generation, ranking, and decision-making. However, their outputs can be influenced by various forms of biases. One such bias is positional bias, where models prioritize items based on their position within a given prompt rather than their content or quality, impacting on how LLMs interpret and weigh information, potentially compromising fairness, reliability, and robustness. To assess positional bias, we prompt a range of LLMs to generate descriptions for a list of topics, systematically permuting their order and analyzing variations in the responses. Our analysis shows that ranking position affects structural features and coherence, with some LLMs also reordering or omitting topics. Nonetheless, the impact of positional bias varies across different LLMs and topics, indicating an interplay with other related biases.
Alignment of large language models (LLMs) with principles like helpfulness, honesty, and harmlessness typically relies on scalar rewards that obscure which objectives drive the training signal. We introduce QA-LIGN, which decomposes monolithic rewards into interpretable principle-specific evaluations through structured natural language programs. Models learn through a draft, critique, and revise pipeline, where symbolic evaluation against the rubrics provides transparent feedback for both initial and revised responses during GRPO training. Applied to uncensored Llama-3.1-8B-Instruct, QA-LIGN reduces attack success rates by up to 68.7% while maintaining a 0.67% false refusal rate, achieving Pareto optimal safety-helpfulness performance and outperforming both DPO and GRPO with state-of-the-art reward models given equivalent training. These results demonstrate that making reward signals interpretable and modular improves alignment effectiveness, suggesting transparency enhances LLM safety.
ProtoXTM: Cross-Lingual Topic Modeling with Document-Level Prototype-based Contrastive Learning
PDF ↗Cross-lingual topic modeling (CLTM) is an essential task in the field of data mining and natural language processing, aiming to extract aligned and semantically coherent topics from bilingual corpora. Recent advances in cross-lingual neural topic models have widely leveraged bilingual dictionaries to achieve word-level topic alignment. However, two critical challenges remain in cross-lingual topic modeling, the topic mismatch issue and the degeneration of intra-lingual topic interpretability. Due to linguistic diversity, some translated word pairs may not represent semantically coherent topics despite being lexical equivalents, and the objective of cross-lingual topic alignment in CLTM can consequently degrade topic interpretability within intra languages. To address these issues, we propose a novel document-level prototype-based contrastive learning paradigm for cross-lingual topic modeling. Additionally, we design a retrieval-based positive sampling strategy for contrastive learning without data augmentation. Furthermore, we introduce ProtoXTM, a cross-lingual neural topic model based on document-level prototype-based contrastive learning. Extensive experiments indicate that our approach achieves state-of-the-art performance on cross-lingual and mono-lingual benchmarks, demonstrating enhanced topic interpretability.
Abusive language online creates toxic environments and exacerbates social tensions, underscoring the need for robust NLP models to interpret nuanced linguistic cues. This paper introduces GAttention, a novel Gated Attention mechanism that combines the strengths of Contextual attention and Self-attention mechanisms to address the limitations of existing attention models within the text classification task. GAttention capitalizes on local and global query vectors by integrating the internal relationships within a sequence (Self-attention) and the global relationships among distinct sequences (Contextual attention). This combination allows for a more nuanced understanding and processing of sequence elements, which is particularly beneficial in context-sensitive text classification tasks such as the case of abusive language detection. By applying this mechanism to transformer-based encoder models, we showcase how it enhances the model’s ability to discern subtle nuances and contextual clues essential for identifying abusive language, a challenging and increasingly relevant NLP task.
Vision-Language Models (VLMs) have demonstrated remarkable capabilities in interpreting visual layouts and text. However, a significant challenge remains in their ability to interpret robustly and reason over multi-tabular data presented as images, a common occurrence in real-world scenarios like web pages and digital documents. Existing benchmarks typically address single tables or non-visual data (text/structured). This leaves a critical gap: they don’t assess the ability to parse diverse table images, correlate information across them, and perform multi-hop reasoning on the combined visual data. To bridge this evaluation gap, we introduce MTabVQA, a novel benchmark specifically designed for multi-tabular visual question answering. MTabVQA comprises 3,745 complex question-answer pairs that necessitate multi-hop reasoning across several visually rendered table images. We provide extensive benchmark results for state-of-the-art VLMs on MTabVQA, revealing significant performance limitations. We further investigate post-training techniques to enhance these reasoning abilities and release MTabVQA-Instruct, a large-scale instruction-tuning dataset. Our experiments show that fine-tuning VLMs with MTabVQA-Instruct substantially improves their performance on visual multi-tabular reasoning. Code and dataset are available online: .
Sparse autoencoders (SAEs) are emerging as a key analytical tool in the field of mechanistic interpretability for large language models (LLMs). While SAEs have primarily been used for interpretability, we shift focus and explore an understudied question: “Can SAEs be applied to practical tasks beyond interpretability?” Given that SAEs are trained on billions of tokens for sparse reconstruction, we believe they can serve as effective extractors, offering a wide range of useful knowledge that can benefit practical applications. Building on this motivation, we demonstrate that SAEs can be effectively applied to in-context learning (ICL). In particular, we highlight the utility of the SAE-reconstruction loss by showing that it provides a valuable signal in ICL—exhibiting a strong correlation with LLM performance and offering a powerful unsupervised approach for prompt selection. These findings underscore the versatility of SAEs and reveal their potential for real-world applications beyond interpretability. Our code is available at https://github.com/ihcho2/SAE-GPS.
ClueAnchor: Clue-Anchored Knowledge Reasoning Exploration and Optimization for Retrieval-Augmented Generation
PDF ↗Retrieval-Augmented Generation (RAG) augments Large Language Models (LLMs) with external knowledge to improve factuality. However, existing RAG systems frequently underutilize the retrieved documents, failing to extract and integrate the key clues needed to support faithful and interpretable reasoning, especially in cases where relevant evidence is implicit, scattered, or obscured by noise. To address this issue, we propose ClueAnchor, a novel framework for enhancing RAG via clue-anchored reasoning exploration and optimization. ClueAnchor extracts key clues from retrieved content and generates multiple reasoning paths based on different knowledge configurations, optimizing the model by selecting the most appropriate reasoning path for the given context through reward-based preference optimization. Experiments show that ClueAnchor significantly outperforms prior RAG baselines in the completeness and robustness of reasoning. Further analysis confirms its strong resilience to noisy or partially relevant retrieved content, as well as its capability to identify supporting evidence even in the absence of explicit clue supervision during inference. All codes are available at https://github.com/thunlp/ClueAnchor.
Temporal knowledge graph (TKG) reasoning, a central task in temporal knowledge representation, focuses on predicting future facts by leveraging historical temporal contexts. However, current approaches face two major challenges: limited generalization to unseen facts and insufficient interpretability of reasoning processes. To address these challenges, this paper proposes the **D**enoising **L**ogic-based **T**emporal **K**nowledge **G**raph (DLTKG) framework, which employs a denoising diffusion process to complete reasoning tasks by introducing a noise source and a historical conditionguiding mechanism. Specifically, DLTKG constructs fuzzy entity representations by treating historical facts as noise sources, thereby enhancing the semantic associations between entities and the generalization ability for unseen facts. Additionally, the condition-based guidance mechanism, rooted in the relationship evolutionary paths, is designed to improve the interpretability of the reasoning process. Furthermore, we introduce a fine-tuning strategy that optimizes the denoising process by leveraging shortest path information between the head entity and candidate entities. Experimental results on three benchmark datasets demonstrate that DLTKG outperforms state-of-the-art methods across multiple evaluation metrics. Our code is available at: https://github.com/NEU-IDKE/DLTKG
From n-gram to Attention: How Model Architectures Learn and Propagate Bias in Language Modeling
PDF ↗Current research on bias in language models (LMs) predominantly focuses on data quality, with significantly less attention paid to model architecture and temporal influences of data. Even more critically, few studies systematically investigate the origins of bias. We propose a methodology grounded in comparative behavioral theory to interpret the complex interaction between training data and model architecture in bias propagation during language modeling. Building on recent work that relates transformers to n-gram LMs, we evaluate how data, model design choices, and temporal dynamics affect bias propagation. Our findings reveal that: (1) n-gram LMs are highly sensitive to context window size in bias propagation, while transformers demonstrate architectural robustness; (2) the temporal provenance of training data significantly affects bias; and (3) different model architectures respond differentially to controlled bias injection, with certain biases (e.g. sexual orientation) being disproportionately amplified. As language models become ubiquitous, our findings highlight the need for a holistic approach- tracing bias to its origins across both data and model dimensions, not just symptoms, to mitigate harm.
Glitter: A Multi-Sentence, Multi-Reference Benchmark for Gender-Fair German Machine Translation
PDF ↗Machine translation (MT) research addressing gender inclusivity has gained attention for promoting non-exclusionary language representing all genders. However, existing resources are limited in size, most often consisting of single sentences, or single gender-fair formulation types, leaving questions about MT models’ ability to use context and diverse inclusive forms. We introduce Glitter, an English-German benchmark featuring extended passages with professional translations implementing three gender-fair alternatives: neutral rewording, typographical solutions (gender star), and neologistic forms (-ens forms). Our experiments reveal significant limitations in state-of-the-art language models, which default to masculine generics, struggle to interpret explicit gender cues in context, and rarely produce gender-fair translations. Through a systematic prompting analysis designed to elicit fair language, we demonstrate that these limitations stem from models’ fundamental misunderstanding of gender phenomena, as they fail to implement inclusive forms even when explicitly instructed. Glitter establishes a challenging benchmark, advancing research in gender-fair English-German MT. It highlights substantial room for improvement among leading models and can guide the development of future MT models capable of accurately representing gender diversity.
Vision-Language Models (VLMs) excel at tasks such as image captioning and visual question answering but frequently produce hallucinated outputs that deviate from the actual visual input or prompt. While prior work links hallucination to biases in data or representation, their causal origins remain unclear. We propose a causal framework to analyze and mitigate hallucination in VLMs. Our key hypothesis is that hallucinations arise from unintended direct influences of the vision or text modality that bypass the intended multi-modal fusion. To examine this, we construct a causal graph of the VLM and use counterfactual analysis to estimate the Natural Direct Effect (NDE) of each modality and their interaction. By systematically identifying and suppressing these direct effects, we encourage outputs that are more faithfully grounded in true cross-modal reasoning. Our approach consists of three steps: (1) designing structural causal graphs to distinguish correct fusion pathways from spurious modality shortcuts, (2) estimating modality-specific and cross-modal NDE using perturbed image representations, hallucinated text embeddings, and degraded visual inputs, and (3) implementing a test-time intervention module to dynamically adjust the model’s dependence on each modality. Experimental results demonstrate that our method significantly reduces hallucination while preserving task performance, providing a robust and interpretable framework for improving VLM reliability.
Adaptive Platt Scaling with Causal Interpretations for Self-Reflective Language Model Uncertainty Estimates
PDF ↗As large language models (LLMs) are consumed by more users and deployed in increasingly autonomous capacities, their ability to self-monitor and ask for human intervention is of vital importance. Underlying this capability are fundamental skills like self-reflection and expression of uncertainty. In this work, we provide a formal analysis of LLM self-reflection for uncertainty estimation, using domain adaptation theory to model the shift between base predictions and reflective judgments. We use this to motivate a temperature scaling algorithm that calibrates uncertainty using comparisons between base predictions and LLM self-reflections. We evaluate our approach on challenging question-answering tasks requiring reasoning, demonstrating that our methods can improve calibration of uncertainty estimates and also offer improvements in human interpretation. More broadly, this use case shows how domain adaptation presents a promising analytical tool for understanding the underlying statistical properties of LLM self-reflections.
Large language models (LLMs) have advanced the automation of data science workflows. Yet it remains unclear whether they can critically leverage external domain knowledge as human data scientists do in practice. To answer this question, we introduce AssistedDS (Assisted Data Science), a benchmark designed to systematically evaluate how LLMs handle domain knowledge in tabular prediction tasks. AssistedDS features both synthetic datasets with explicitly known generative mechanisms and real-world Kaggle competitions, each accompanied by curated bundles of helpful and adversarial documents. These documents provide domain-specific insights into data cleaning, feature engineering, and model selection. We assess state-of-the-art LLMs on their ability to discern and apply beneficial versus harmful domain knowledge, evaluating submission validity, information recall, and predictive performance. Our results demonstrate three key findings: (1) LLMs frequently exhibit an uncritical adoption of provided information, significantly impairing their predictive performance when adversarial content is introduced, (2) helpful guidance is often insufficient to counteract the negative influence of adversarial information, and (3) in Kaggle datasets, LLMs often make errors in handling time-series data, applying consistent feature engineering across different folds, and interpreting categorical variables correctly. These findings highlight a substantial gap in current models’ ability to critically evaluate and leverage expert knowledge, underscoring an essential research direction for developing more robust, knowledge-aware automated data science systems. Our data and code are publicly available [here](https://github.com/jeremyxianx/Assisted-DS).
Language models can be steered by modifying their internal representations to control concepts such as emotion, style, or truthfulness in generation. However, the conditions for an effective intervention remain unclear and are often validated through heuristics and trial-and-error. To fill this gap, we demonstrate that intervention efficacy, measured by linear steerability (i.e., the ability to adjust output via linear transformations of hidden states), emerges during intermediate stages of training. Moreover, even closely related concepts (e.g., anger and sadness) exhibit steerability emergence at distinct stages of training*.To better interpret the dynamics of steerability during training, we adapt existing intervention techniques into a unified framework, referred to as the “Intervention Detector” (ID), which is designed to reveal how linear steerability evolves over the course of training through hidden state and representation analysis. ID reveals that concepts become increasingly linearly separable in the hidden space as training progresses, which strongly correlates with the emergence of linear steerability. We further introduce ID-based metrics, such as heatmaps, entropy trends, and cosine similarity, to help interpret how linear steerability evolves throughout training. In addition, we apply ID across different model families to ensure the generality of our findings on steerability dynamics.
Effective query expansion for web search benefits from promoting both exploration and result diversity to capture multiple interpretations and facets of a query. While recent LLM-based methods have improved retrieval performance and demonstrate strong domain generalization without additional training, they often generate narrowly focused expansions that overlook these desiderata. We propose ThinkQE, a test-time query expansion framework addressing this limitation through two key components: a thinking-based expansion process that encourages deeper and comprehensive semantic exploration, and a corpus-interaction strategy that iteratively refines expansions using retrieval feedback from the corpus. Experiments on diverse web search benchmarks (DL19, DL20, and BRIGHT) show ThinkQE consistently outperforms prior approaches, including training-intensive dense retrievers and rerankers.