Zhen Yang, Wei Du, Jie Wang, Wenze Zhou, Xiangfeng Meng, Zhengyang Wang, Suping Sun, Ziwei Du, Haodong Zou, Jie Chen, Yongbin Liu, Shicheng Tan, Jiahao Ying, Shu Zhao. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Wei Tang, Yixin Cao, Yang Deng, Jiahao Ying, Bo Wang, Yizhe Yang, Yuyue Zhao, Qi Zhang, Xuanjing Huang, Yu-Gang Jiang, Yong Liao. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

This paper introduces a Dual Evaluation Framework to comprehensively assess the multilingual capabilities of LLMs.By decomposing the evaluation along the dimensions of linguistic medium and cultural context, this framework enables a nuanced analysis of LLMs' ability to process questions within both native and cross-cultural contexts cross-lingually.Extensive evaluations are conducted on a wide range of models, revealing a notable "Cultural-Linguistic Synergy" phenomenon, where models exhibit better performance when questions are culturally aligned with the language.This phenomenon is further explored through interpretability probing, which shows that a higher proportion of specific neurons are activated in a language's cultural context.This activation proportion could serve as a potential indicator for evaluating multilingual performance during model training.Our findings challenge the prevailing notion that LLMs, primarily trained on English data, perform uniformly across languages and highlight the necessity of culturally and linguistically model evaluations.Our code can be found at https://yingjiahao14. github.io/Dual-Evaluation/.

This study investigates the behaviors of Large Language Models (LLMs) when faced with conflicting prompts versus their internal memory.This will not only help to understand LLMs' decision mechanism but also benefit real-world applications, such as retrievalaugmented generation (RAG).Drawing on cognitive theory, we target the first scenario of decision-making styles where there is no superiority in the conflict and categorize LLMs' preference into dependent, intuitive, and rational/irrational styles.Another scenario of factual robustness considers the correctness of prompt and memory in knowledge-intensive tasks, which can also distinguish if LLMs behave rationally or irrationally in the first scenario.To quantify them, we establish a complete benchmarking framework including a dataset, a robustness evaluation pipeline, and corresponding metrics.Extensive experiments with seven LLMs reveal their varying behaviors.And, with role play intervention, we can change the styles, but different models present distinct adaptivity and upper-bound.One of our key takeaways is to optimize models or the prompts according to the identified style.For instance, RAG models with high role play adaptability may dynamically adjust the interventions according to the quality of retrieval results -being dependent to better leverage informative context; and, being intuitive when the external prompt is noisy.Our dataset can be found at https://github.com/yingjiahao14/KRE.

Large language models (LLMs) have achieved impressive performance across various natural language benchmarks, prompting a continual need to curate more difficult datasets for larger LLMs, which is costly and time-consuming. In this paper, we propose to automate dataset updating and provide systematical analysis regarding its effectiveness in dealing with benchmark leakage issue, difficulty control, and stability. Thus, once current benchmark has been mastered or leaked, we can update it for timely and reliable evaluation. There are two updating strategies: 1) mimicking strategy to generate similar samples based on original data, preserving stylistic and contextual essence, and 2) extending strategy that further expands existing samples at varying cognitive levels by adapting Bloom’s taxonomy of educational objectives. Extensive experiments on updated MMLU and BIG-Bench demonstrate the stability of the proposed strategies and find that the mimicking strategy can effectively alleviate issues of overestimation from benchmark leakage. In cases where the efficient mimicking strategy fails, our extending strategy still shows promising results. Additionally, by controlling the difficulty, we can better discern the models’ performance and enable fine-grained analysis — neither too difficult nor too easy an exam can fairly judge students’ learning status. To the best of our knowledge, we are the first to automate updating benchmarks for reliable and timely evaluation. Our demo leaderboard can be found at https://yingjiahao14.github.io/Automating-DatasetUpdates/.

Numerous benchmarks have been established to assess the performance of foundation models on open-ended question answering, which serves as a comprehensive test of a model's ability to understand and generate language in a manner similar to humans.Most of these works focus on proposing new datasets, however, we see two main issues within previous benchmarking pipelines, namely testing leakage and evaluation automation. In this paper, we propose a novel benchmarking framework, Language-Model-as-an-Examiner, where the LM serves as a knowledgeable examiner that formulates questions based on its knowledge and evaluates responses in a reference-free manner. Our framework allows for effortless extensibility as various LMs can be adopted as the examiner, and the questions can be constantly updated given more diverse trigger topics. For a more comprehensive and equitable evaluation, we devise three strategies: (1) We instruct the LM examiner to generate questions across a multitude of domains to probe for a broad acquisition, and raise follow-up questions to engage in a more in-depth assessment. (2) Upon evaluation, the examiner combines both scoring and ranking measurements, providing a reliable result as it aligns closely with human annotations. (3) We additionally propose a decentralized Peer-examination method to address the biases in a single examiner. Our data and benchmarking results are available at: http://lmexam.xlore.cn.

Multiple kernel clustering (MKC) optimally combines a group of predefined kernel matrices to improve clustering performance. Although demonstrating promising performance in various applications, most of existing approaches adopt the min–min formulation, which could be sensitive to perturbation with adversarial samples. Moreover, existing MKC algorithms often involve several hypermeters preventing them into further real applications. To address these issues, we propose a parameter-free effective sparse simple multiple kernel k-means algorithm with max–min optimization formulation in this paper. To be specific, we propose to optimize the widely used unsupervised kernel alignment criterion by minimizing the kernel coefficient and maximizing the clustering partition matrix. Unlike traditional min–min formulation, the max–min kernel alignment is robust to adversarial sample perturbation and free of hyper-parameters. An optimization method based on semi-infinite linear program is designed to solve the complicated optimization problem. Extensive experiments on six multiple kernel benchmark data sets demonstrate the effectiveness of the proposed method.