1077 papers
This study demonstrates that integrating coreference resolution into Retrieval-Augmented Generation (RAG) systems significantly enhances retrieval relevance and question-answering performance, particularly for smaller language models, by resolving entity ambiguities that otherwise disrupt contextual understanding.
This paper employs interpretability techniques on the off-by-one addition task to reveal that large language models achieve task-level generalization through a reusable "function induction" mechanism, where multiple attention heads collaboratively learn and compose abstract functions to solve unseen problems.
This paper introduces the first comprehensive multilingual benchmark for evaluating text detoxification across nine languages, demonstrating that proposed neural-based and LLM-as-a-judge metrics significantly outperform existing baselines in correlating with human judgments.
The paper introduces Text2VLM, a novel pipeline that converts text-only datasets into multimodal prompts to evaluate and reveal the heightened susceptibility of open-source Visual Language Models to typographic prompt injection attacks, thereby advancing the development of robust safety mechanisms for multimodal AI.
This paper introduces WebDS, the first end-to-end benchmark for web-based data science comprising 870 complex, multi-step tasks across diverse websites, which reveals significant performance gaps between current LLM agents and human capabilities in real-world data acquisition, analysis, and insight generation.
The paper introduces SEVADE, a novel self-evolving multi-agent framework featuring a Dynamic Agentive Reasoning Engine and a decoupled adjudicator to enhance sarcasm detection accuracy and reliability by mitigating hallucinations through structured, multifaceted reasoning.
This paper introduces ObfusQAte and the ObfusQA framework, a novel benchmark designed to evaluate Large Language Model robustness by testing their performance on factual questions subjected to multi-tiered obfuscation, revealing that models often fail or hallucinate when faced with nuanced linguistic variations.
The paper introduces MultiWikiQA, a large-scale reading comprehension benchmark covering 306 languages with over 1.2 million samples generated from Wikipedia articles and validated by human evaluation, demonstrating significant performance disparities across languages and model architectures.
This paper proposes a lightweight token pruning framework that filters non-informative background regions and refines fragmented text areas in document images to significantly reduce computational costs in vision-language models while maintaining comparable accuracy.
This paper addresses the unreliability of multimodal agents in executing toggle control instructions within GUIs by proposing State-aware Reasoning (StaR), a method that significantly improves toggle execution accuracy by over 30% through perceiving current states and inferring desired outcomes.
This paper introduces the Contextualized Argument Appraisal Framework, which integrates psychological appraisal models to analyze how subjective factors like familiarity and emotional response influence argument convincingness, validated through a novel role-playing annotation setup that produced the 4,000-entry ContArgA corpus.
This paper introduces a novel user simulator that generates four categories of non-collaborative behaviors to expose significant robustness weaknesses in current tool agents, thereby highlighting the critical need for training agents against realistic, challenging user interactions.
This paper introduces PDR-Bench, the first benchmark for evaluating personalization in Deep Research Agents, along with the PQR Evaluation Framework to assess performance across personalized, open-ended research tasks.
The paper introduces GraphMERT, a compact, efficient encoder-only model that overcomes the scalability and reliability limitations of large language models by distilling high-quality, factually accurate, and ontology-consistent knowledge graphs from unstructured text, thereby establishing a scalable neurosymbolic framework that significantly outperforms existing baselines in both accuracy and symbolic validity.
This paper proposes a novel geometric framework that models LLM reasoning as smooth flows in representation space, demonstrating through empirical experiments that next-token prediction enables models to internalize logical invariants as higher-order geometry, thereby challenging the "stochastic parrot" hypothesis and suggesting a universal representational law underlying machine understanding.
This paper introduces WeightLens and CircuitLens, two complementary methods that advance mechanistic interpretability by analyzing feature weights and component interactions directly, thereby overcoming the limitations of activation-based approaches in scalability, robustness, and the ability to capture circuit-level dynamics without relying on external explainer models or datasets.
This paper introduces COGS, a data-efficient framework that synthesizes large-scale reasoning datasets by decomposing seed questions into primitive factors and recomposing them with new images, thereby significantly enhancing the visual reasoning capabilities of multi-modal large language models in annotation-scarce domains like charts and webpages.
This paper introduces EliCal, an annotation-efficient two-stage framework that combines inexpensive self-consistency elicitation with minimal correctness labeling to achieve universal honesty alignment in large language models, validated by the newly released HonestyBench benchmark.
This paper introduces the CITECONTROL benchmark to systematically analyze citation failures in LLM-based RAG systems and proposes CITENTION, a hybrid framework that effectively mitigates these failures by integrating generative, attention-based, and retrieval-based methods.
The paper proposes REVISION, a novel framework that integrates offline large-model reasoning to mine implicit user intents from no-click requests with online adaptive decision-making to optimize Taobao's visual search system and significantly reduce no-click rates.