This study proposes a simple yet effective LoRA merge method to achieve LLM adaptation for low-resource language generation tasks. The LoRA merge technique, which integrates multiple LoRA modules trained on different tasks, has gained attention as an effective and efficient approach for adapting LLMs to target tasks. However, previous methods are limited in adaptability as they keep the LoRA parameters frozen. Additionally, the low-resource problem has been out of their scope. We propose a LoRA merge method that updates and prunes LoRA parameters through fine-tuning with minimal target task data, which allows finer-grained adjustments of LoRA parameters and enhancement of task adaptability. Extensive experiments have been conducted taking summarization as a benchmark task. Our datasets cover various domains and multiple languages of English and Japanese. The results confirm that the proposed method achieves significant and consistent improvements in task adaptability over the previous methods.
论文检索
输入标题、作者或关键词，从 12,319 篇学术成果中精准定位
Recent works have highlighted the significance of memory mechanisms in LLM-based agents, which enable them to store observed information and adapt to dynamic environments. However, evaluating their memory capabilities still remains challenges. Previous evaluations are commonly limited by the diversity of memory levels and interactive scenarios. They also lack comprehensive metrics to reflect the memory capabilities from multiple aspects. To address these problems, in this paper, we construct a more comprehensive dataset and benchmark to evaluate the memory capability of LLM-based agents. Our dataset incorporates factual memory and reflective memory as different levels, and proposes participation and observation as various interactive scenarios. Based on our dataset, we present a benchmark, named MemBench, to evaluate the memory capability of LLM-based agents from multiple aspects, including their effectiveness, efficiency, and capacity. To benefit the research community, we release our dataset and project at https://github.com/import-myself/Membench.
Human activity is moderated by norms; however, supervision for normative reasoning is sparse, particularly where norms are physically- or socially-grounded. We thus present EgoNormia \lVert \epsilon \rVert, comprising 1,853 (200 for EgoNormia-verified) multiple choice questions (MCQs) grounded within ego-centric videos of human interactions, enabling the evaluation and improvement of normative reasoning in vision-language models (VLMs). spans seven norm categories: safety, privacy, proxemics, politeness, cooperation, coordination/proactivity, and communication/legibility. To compile this dataset at scale, we propose a novel pipeline to generate grounded MCQs from raw egocentric video. Our work demonstrates that current state-of-the-art VLMs lack robust grounded norm understanding, scoring a maximum of 54% on EgoNormia and 58% on EgoNormia-verified, with performance across norm categories indicating significant risks of safety and privacy when VLMs are used in real-world agents. We additionally explore methods for improving normative understanding, demonstrating a naive retrieval-based generation (RAG) method using can enhance normative reasoning in VLMs.
This study intends to systematically disentangle pure logic reasoning and text understanding by investigating the contrast across abstract and contextualized logical problems from a comprehensive set of domains. We explore whether LLMs demonstrate genuine reasoning capabilities across various domains when the underlying logical structure remains constant. We focus on two main questions (1) Can abstract logical problems alone accurately benchmark LLMs’ reasoning ability in real-world scenarios, disentangled from contextual support in practical settings? (2) Does fine-tuning LLMs on abstract logic problems generalize to contextualized logic problems and vice versa? To investigate these questions, we focus on standard propositional logic, specifically propositional deductive and abductive logic reasoning. We construct datasets for both reasoning types with four difficulty levels across 12 distinct domains based on the Wikipedia categorization in addition to those with purely abstract variables. Our experiments aim to provide insights into disentangling context in logical reasoning, the genuine reasoning capabilities of LLMs, and their generalization potential. Coda and data are available at https://anonymous.4open.science/r/ContextHub-957E.
Large-scale multilingual evaluations, such as MEGA, often include only a handful of African languages due to the scarcity of high-qualityevaluation data and the limited discoverability of existing African datasets. This lack of representation hinders comprehensive LLM evaluation across a diverse range of languages and tasks. To address these challenges, we introduce AFROBENCH—a multi-task benchmark for evaluating the performance of LLMs across 64 African languages, 15 tasks and 22 datasets. AFROBENCH consists of nine natural language understanding datasets, six text generation datasets, six knowledge and question answering tasks, and one mathematical reasoning task. We present results comparing the performance of prompting LLMs to fine-tuned baselines based on BERT and T5-style models. Our results suggest large gaps in performance between high-resource languages, such as English, and African languages across most tasks; but performance also varies based on the availability of monolingual data resources. Our findings confirm that performance on African languages continues to remain a hurdle for current LLMs, underscoring the need for additional efforts to close this gap.
A Multi-Labeled Dataset for Indonesian Discourse: Examining Toxicity, Polarization, and Demographics Information
PDF ↗Online discourse is increasingly trapped in a vicious cycle where polarizing language fuelstoxicity and vice versa. Identity, one of the most divisive issues in modern politics, oftenincreases polarization. Yet, prior NLP research has mostly treated toxicity and polarization asseparate problems. In Indonesia, the world’s third-largest democracy, this dynamic threatens democratic discourse, particularly in online spaces. We argue that polarization and toxicity must be studied in relation to each other. To this end, we present a novel multi-label Indonesian dataset annotated for toxicity, polarization, and annotator demographic information. Benchmarking with BERT-base models and large language models (LLMs) reveals that polarization cues improve toxicity classification and vice versa. Including demographic context further enhances polarization classification performance.
In the last couple years, there has been a flood of interest in studying the extent to which language models (LMs) have a theory of mind (ToM) — the ability to ascribe mental states to themselves and others. The results provide an unclear picture of the current state of the art, with some finding near-human performance and others near-zero. To make sense of this landscape, we perform a survey of 16 recent studies aimed at measuring ToM in LMs and find that, while almost all perform checks for human identifiable issues, less than half do so for patterns only a machine might exploit. Among those that do perform such validation, which we call machine validation, none identify LMs to exceed human performance. We conclude that the datasets that show high LM performance on ToM tasks are easier than their peers, likely due to the presence of spurious patterns in the data, and we caution against building ToM benchmarks relying solely on human validation of the data.
Post-earnings volatility prediction is critical for investors, with previous works often leveraging earnings call transcripts under the assumption that their rich semantics contribute significantly. To further investigate how transcripts impact volatility, we introduce DEC, a dataset featuring accurate volatility calculations enabled by the previously overlooked beforeAfterMarket attribute and dense ticker coverage. Unlike established benchmarks, where each ticker has only around two earnings, DEC provides 20 earnings records per ticker. Using DEC, we reveal that post-earnings volatility undergoes significant shifts, with each ticker displaying a distinct volatility distribution. To leverage historical post-earnings volatility and capture ticker-specific patterns, we propose two training-free baselines: Post-earnings Volatility (PEV) and Same-ticker Post-earnings Volatility (STPEV). These baselines surpass all transcripts-based models on DEC as well as on established benchmarks. Additionally, we demonstrate that current transcript representations predominantly capture ticker identity rather than offering financially meaningful insights specific to each earnings. This is evidenced by two key observations: earnings representations from the same ticker exhibit significantly higher similarity compared to those from different tickers, and predictions from transcript-based models show strong correlations with prior post-earnings volatility.
Ideation, the process of forming ideas from concepts, is a big part of the content creation process. However, the noble goal of helping visual content creators by suggesting meaningful sequences of visual assets from a limited collection is challenging. It requires a nuanced understanding of visual assets and the integration of open-world knowledge to support creative exploration. Despite its importance, this task has yet to be explored fully in existing literature. To fill this gap, we propose Visual Story Ideation, a novel and underexplored task focused on the automated selection and arrangement of visual assets into coherent sequences that convey expressive storylines.We also present VISIAR, Visual Ideation through Sequence Integration and Asset Rearrangement, a robust framework leveraging Multimodal Large Language Models (MLLMs), and a novel Story Graph mechanism. Our framework operates in three key stages: visual content understanding, candidate asset selection, and asset rearrangement via MLLMs. In addition, we curated a new benchmark dataset, called VTravel, to evaluate our methods both qualitatively and quantitatively.User studies and GPT-as-the-judge evaluation show that our approach surpasses GPT-4o based baseline by an average of 33.5% and 18.5% across three different metrics, demonstrating the effectiveness of our framework for generating compelling visual stories.
PromptCoT: Synthesizing Olympiad-level Problems for Mathematical Reasoning in Large Language Models
PDF ↗The ability of large language models to solve complex mathematical problems has progressed significantly, particularly for tasks requiring advanced reasoning. However, the scarcity of sufficiently challenging problems, particularly at the Olympiad level, hinders further advancements. In this work, we introduce PromptCoT, a novel approach for automatically generating high-quality Olympiad-level math problems. The proposed method synthesizes complex problems based on mathematical concepts and the rationale behind problem construction, emulating the thought processes of experienced problem designers. We provide a theoretical analysis demonstrating that an optimal rationale should maximize both the likelihood of rationale generation given the associated concepts and the likelihood of problem generation conditioned on both the rationale and the concepts. Our method is evaluated on standard benchmarks including GSM8K, MATH-500, and AIME2024, where it consistently outperforms existing problem generation methods. Furthermore, we demonstrate that PromptCoT exhibits superior scalability, consistently maintaining high performance as the dataset size increases, outperforming the baselines.
Significant efforts have been made in the NLP community to facilitate the automatic analysis of climate-related corpora by tasks such as climate-related topic detection, climate risk classification, question answering over climate topics, and many more. In this work, we perform a reproducibility study on 8 tasks and 29 datasets, testing 6 models. We find that many tasks rely heavily on surface-level keyword patterns rather than deeper semantic or contextual understanding. Moreover, we find that 96% of the datasets contain annotation issues, with 16.6% of the sampled wrong predictions of a zero-shot classifier being actually clear annotation mistakes, and 38.8% being ambiguous examples.These results call into question the reliability of current benchmarks to meaningfully compare models and highlight the need for improved annotation practices. We conclude by outlining actionable recommendations to enhance dataset quality and evaluation robustness.
Despite recent advances in Video Large Language Models (VideoLLMs), effectively understanding long-form videos remains a significant challenge. Perceiving lengthy videos containing thousands of frames poses substantial computational burden. To mitigate this issue, this paper introduces Generative Frame Sampler (GenS), a plug-and-play module integrated with VideoLLMs to facilitate efficient lengthy video perception. Built upon a lightweight VideoLLM, GenS leverages its inherent vision-language capabilities to identify question-relevant frames. To facilitate effective retrieval, we construct GenS-Video-150K, a large-scale video instruction dataset with dense frame relevance annotations. Extensive experiments demonstrate that GenS consistently boosts the performance of various VideoLLMs, including open-source models (Qwen2-VL-7B, Aria-25B, LLaVA-Video-7B/72B) and proprietary assistants (GPT-4o, Gemini). When equipped with GenS, open-source VideoLLMs achieve impressive state-of-the-art results on long-form video benchmarks: LLaVA-Video-72B reaches 66.8 (+4.3) on LongVideoBench and 77.0 (+2.7) on MLVU, while Aria obtains 39.2 on HourVideo surpassing the Gemini-1.5-pro by 1.9 points.
100-LongBench: Are de facto Long-Context Benchmarks Literally Evaluating Long-Context Ability?
PDF ↗Long-context capability is considered one of the most important abilities of LLMs, as a truly long context-capable LLM shall enable its users to effortlessly process many originally exhausting tasks — e.g., digesting a long-form document to find answers v.s., directly asking an LLM about it. However, existing real-task-based long-context evaluation benchmarks have a few major shortcomings. For instance, some Needle-in-a-Haystack-like benchmarks are too synthetic, and therefore do not represent the real world usage of LLMs. While some real-task-based benchmarks like LongBench avoid this problem, such benchmarks are often formed in a way where each data sample has a fixed sequence length, which not only makes them solely suitable for models with a certain range of context windows, but also lacks a proxy to know at what length the model/method-of-interest would fail. Last, most benchmarks tend to not provide proper metrics to separate long-context performance from the model’s baseline ability, so when conducting a cross-model/recipe comparison, such conflation makes the user unable to understand how exactly one model or recipe excels at the long-context task in relation to its baseline ability. To address these issues, we introduce a length-controllable, real-life reflective benchmark with a novel metric that disentangles baseline knowledge from long-context capabilities. Experiments demonstrate the superiority of our datasets in effectively evaluating LLMs. All assets are available at https://github.com/uservan/100-LongBench.git.
This paper investigates the effect of dialectal prompting, variations in prompting scrip t and model fine-tuning on subjective classification in Arabic dialects. To this end, we evaluate the performances of 12 widely used open LLMs across four tasks and eight benchmark datasets. Our results reveal that specialized fine-tuned models with Arabic and Arabizi scripts dialectal prompts achieve the best results, which constitutes a novel state of the art in the field.
Radiology report generation (RRG) requires advanced medical image analysis, effective temporal reasoning, and accurate text generation. While multimodal large language models (MLLMs) align with pre-trained vision encoders to enhance visual-language understanding, most existing methods rely on single-image analysis or rule-based heuristics to process multiple images, failing to fully leverage temporal information in multi-modal medical datasets. In this paper, we introduce **Libra**, a temporal-aware MLLM tailored for chest X-ray report generation. Libra combines a radiology-specific image encoder with a novel Temporal Alignment Connector (**TAC**), designed to accurately capture and integrate temporal differences between paired current and prior images. Extensive experiments on the MIMIC-CXR dataset demonstrate that Libra establishes a new state-of-the-art benchmark among similarly scaled MLLMs, setting new standards in both clinical relevance and lexical accuracy. All source code and data are publicly available at: https://github.com/X-iZhang/Libra.
A Law Reasoning Benchmark for LLM with Tree-Organized Structures including Factum Probandum, Evidence and Experiences
PDF ↗While progress has been made in legal applications, law reasoning, crucial for fair adjudication, remains unexplored. We propose a transparent law reasoning schema enriched with hierarchical factum probandum, evidence, and implicit experience, enabling public scrutiny and preventing bias. Inspired by this schema, we introduce the challenging task, which takes a textual case description and outputs a hierarchical structure justifying the final decision. We also create the first crowd-sourced dataset for this task, enabling comprehensive evaluation. Simultaneously, we propose TL agent that employs a comprehensive suite of legal analysis tools to address the challenge task. This benchmark paves the way for transparent and accountable AI-assisted law-reasoning in the “Intelligent Court”.
Multimodal entity linking (MEL), a task aimed at linking mentions within multimodal contexts to their corresponding entities in a knowledge base (KB), has attracted much attention due to its wide applications in recent years. However, existing MEL methods often rely on mention words as retrieval cues, which limits their ability to effectively utilize information from both images and text. This reliance causes MEL to struggle with accurately retrieving entities in certain scenarios, especially when the focus is on image objects or mention words are missing from the text. To solve these issues, we introduce a Visual Prompts guided Multimodal Entity Linking (VP-MEL) task. Given a text-image pair, VP-MEL aims to link a marked region (i.e., visual prompt) in an image to its corresponding entities in the knowledge base. To facilitate this task, we present a new dataset, VPWiki, specifically designed for VP-MEL. Furthermore, we propose a framework named IIER, which enhances visual feature extraction using visual prompts and leverages the pre-trained Detective-VLM model to capture latent information. Experimental results on the VPWiki dataset demonstrate that IIER outperforms baseline methods across multiple benchmarks for the VP-MEL task.
MAGIC-VQA: Multimodal And Grounded Inference with Commonsense Knowledge for Visual Question Answering
PDF ↗Visual Question Answering (VQA) necessitates models to reason effectively across visual and textual modalities. However, existing Large Vision-Language Models (LVLMs) often fall short in achieving human-like reasoning due to a lack of integrated commonsense knowledge, limiting their robustness and accuracy in real-world scenarios where both explicit facts and implicit understanding are crucial. To address this challenge, we present MAGIC-VQA: Multimodal And Grounded Inference with Commonsense Knowledge, a novel framework designed to enhance multimodal inference by integrating commonsense reasoning. MAGIC-VQA introduces a three-stage process: (1) Explicit Commonsense Knowledge Retrieval from external knowledge graphs, (2) By-Type Commonsense Knowledge Post-Processing to refine contextual relevance, and (3) Implicit Commonsense Knowledge Augmentation using a heterogeneous graph processed by a Graph Neural Network (GNN). These stages collectively enable nuanced, context-aware reasoning without extensive pre-training or intricate prompt tuning.Our MAGIC-VQA significantly improves comprehensive benchmark datasets, surpassing existing models in tasks requiring advanced commonsense reasoning. MAGIC-VQA establishes a robust pathway for integrating commonsense knowledge into VQA, bridging the gap between vision-language inputs and high-level reasoning for improved reliability and contextual accuracy.
Style embeddings are useful for stylistic analysis and style transfer, yet they only exist for English. We introduce Multilingual StyleDistance (mStyleDistance), a method that can generate style embeddings in new languages using synthetic data and a contrastive loss. We create style embeddings in nine languages and a multilingual STEL-or-Content benchmark (Wegmann et al., 2022) that serves to assess their quality. We also employ our embeddings in an authorship verification task involving different languages. Our results show that mStyleDistance embeddings outperform existing style embeddings on these benchmarks and generalize well to unseen features and languages. We make our models and datasets publicly available.
Vulnerability of Text-to-Image Models to Prompt Template Stealing: A Differential Evolution Approach
PDF ↗Prompt trading has emerged as a significant intellectual property concern in recent years, where vendors entice users by showcasing sample images before selling prompt templates that can generate similar images. This work investigates a critical security vulnerability: attackers can steal prompt templates using only a limited number of sample images. To investigate this threat, we introduce \textbf{Prism}, a prompt-stealing benchmark consisting of 50 templates and 450 images, organized into Easy and Hard difficulty levels. To identify the vulnerabity of VLMs to prompt stealing, we propose \textbf{EvoStealer}, a novel template stealing method that operates without model fine-tuning by leveraging differential evolution algorithms. The system first initializes population sets using multimodal large language models (MLLMs) based on predefined patterns, then iteratively generates enhanced offspring through MLLMs. During evolution, EvoStealer identifies common features across offspring to derive generalized templates. Our comprehensive evaluation conducted across open-source (InternVL2-26B) and closed-source models (GPT-4o and GPT-4o-mini) demonstrates that EvoStealer’s stolen templates can reproduce images highly similar to originals and effectively generalize to other subjects, significantly outperforming baseline methods with an average improvement of over 10%. Moreover, our cost analysis reveals that EvoStealer achieves template stealing with negligible computational expenses. Our code and dataset are available at https://whitepagewu.github.io/evostealer-site.