1065 papers
This paper presents a data-driven survey of 14,648 studies from 2022 to early 2025, revealing that research on the limitations of large language models (LLLMs) has surged to over 30% of all LLM-related work, with reasoning, generalization, and hallucination being the most prominent areas of focus.
AutoPCR is a novel, training-free prompt-based framework for phenotype concept recognition that combines rule-based and neural extraction, SapBERT retrieval, and LLM linking to achieve state-of-the-art, robust performance across diverse datasets and ontologies without requiring ontology-specific training.
This paper reveals that the Key-Value (KV) cache used to accelerate Large Language Model inference is vulnerable to privacy attacks that allow attackers to reconstruct sensitive user inputs, and it proposes KV-Cloak, a lightweight and efficient obfuscation defense that effectively prevents such leakage without compromising model accuracy or performance.
The paper introduces LaTeXTrans, a collaborative multi-agent system that overcomes the challenges of translating structured LaTeX documents by decomposing, translating, validating, and reconstructing content to ensure both semantic accuracy and the preservation of complex formatting and syntax.
This paper introduces QCSE, a pretrained quantum context-sensitive word embedding model that utilizes innovative context matrix computation methods to capture linguistic relationships and demonstrates its effectiveness on both English and low-resource Fulani corpora, highlighting the potential of Quantum Natural Language Processing to address data scarcity challenges.
This paper presents a large-scale empirical study across 23 visual question-answering benchmarks that reveals significant variations in intra- and inter-modality dependencies, uncovering that many benchmarks inadvertently amplify image-only reliance while showing limited true multi-modal interaction, thereby proposing a quantitative framework for principled dataset design and evaluation.
The paper proposes Semantic-Anchor Compression (SAC), an autoencoding-free context compression method that improves LLM inference by directly selecting and enhancing specific anchor tokens with learnable embeddings and bidirectional attention to aggregate contextual information, thereby outperforming existing autoencoding-based approaches in question-answering and summarization tasks.
The paper proposes SAGE, a novel user simulation framework for evaluating multi-turn agents that integrates top-down business logic and bottom-up infrastructure knowledge to generate realistic, diverse interactions capable of identifying significantly more agent errors than existing methods.
This paper presents a CEFR-annotated WordNet created using large language models to align semantic definitions with language proficiency levels, demonstrating that the resulting dataset and classifiers effectively bridge natural language processing and language education by achieving performance comparable to gold-standard annotations.
This paper introduces a novel 21-way multiparallel EuroParl dataset to assess political bias in multilingual Large Language Models by demonstrating that translation quality systematically favors majority parties over outsider groups, offering a fairness-based alternative to traditional English survey methods.
KVTC is a lightweight, model-agnostic transform coder that achieves up to 20 (or higher) compression of Key-Value caches for large language models by combining PCA-based decorrelation, adaptive quantization, and entropy coding, thereby enabling memory-efficient serving with reusable caches while maintaining high reasoning and long-context accuracy.
This paper introduces LALITA, a framework that curates efficient parallel corpora for low-resource machine translation by selecting complex source sentences based on lexical and linguistic features, thereby significantly improving translation quality while reducing data requirements by more than half.
The paper introduces V-Skip, a dual-path anchoring mechanism that prevents visual amnesia in multimodal reasoning by reformulating token compression as a visual-anchored information bottleneck, thereby achieving a 2.9× speedup with negligible accuracy loss.
This paper establishes a rate-distortion theorem demonstrating that hallucinations in large language models are an inevitable consequence of information-theoretic optimal memory compression when storing sparse facts, forcing the model to confidently assign high scores to non-facts rather than abstain.
This paper introduces EverMemBench, the first benchmark designed to evaluate long-horizon memory in multi-party collaborative dialogues, revealing that current LLM systems struggle with multi-hop reasoning, temporal versioning, and implicit relevance retrieval in realistic, complex interaction scenarios.
The paper introduces PsihoRo, the first open-source Romanian corpus for depression and anxiety, which was constructed using open-ended responses and standardized screening surveys (PHQ-9 and GAD-7) from 205 participants to address the lack of mental health resources for the Romanian language in NLP.
This paper investigates how Large Language Models trained on the BabyLM dataset often fail to learn specific grammatical rules because early erroneous biases, driven by misleading bigram statistics, become entrenched and persist throughout training, hindering efficient learning.
AdaPonderLM is a self-supervised recurrent language model that employs token-wise adaptive halting gates and KV reuse to dynamically allocate inference compute to difficult tokens, achieving significant efficiency gains without sacrificing performance compared to fixed-depth baselines.
This paper introduces Cross-Family Speculative Prefill, a training-free method that leverages lightweight draft models from different families to compress long prompts for target LLMs, achieving substantial latency reductions while maintaining or slightly improving accuracy across diverse tasks.
The paper introduces GhazalBench, a benchmark demonstrating that while large language models can generally understand the meaning of Persian ghazals, they struggle with exact verse recall in completion tasks due to limited training exposure, a gap that narrows in recognition settings and is less pronounced for English sonnets.