Austria

Commonsense question answering (QA) are widely used to evaluate the commonsense abilities of large language models. However, answering commonsense questions correctly requires not only knowledge but also reasoning—even for seemingly simple questions. We demonstrate that such hidden reasoning attributes in commonsense questions can lead evaluation accuracy differences of up to 24.8% across different difficulty levels in the same benchmark. Current benchmarks overlook these hidden reasoning attributes, making it difficult to assess a model&#39;s specific levels of commonsense knowledge and reasoning ability. To address this issue, we introduce ReComSBench, a novel framework that reveals hidden reasoning attributes behind commonsense questions by leveraging the knowledge generated during the reasoning process. Additionally, ReComSBench proposes three new metrics for decoupled evaluation: Knowledge Balanced Accuracy, Marginal Sampling Gain, and Knowledge Coverage Ratio. Experiments show that ReComSBench provides insights into model performance that traditional benchmarks cannot offer. The difficulty stratification based on revealed hidden reasoning attributes performs as effectively as the model-probability-based approach but is more generalizable and better suited for improving a model&#39;s commonsense reasoning abilities. By uncovering and analyzing the hidden reasoning attributes in commonsense data, ReComSBench offers a new approach to enhancing existing commonsense benchmarks.

ACL 2025

Beyond Surface Simplicity: Revealing Hidden Reasoning Attributes for Precise Commonsense Diagnosis

reasoning

commonsense

benchmark

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Please register!

NAACL 2025

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Commonsense question answering (QA) are widely used to evaluate the commonsense abilities of large language models. However, answering commonsense questions correctly requires not only knowledge but also reasoning—even for seemingly simple questions. We demonstrate that such hidden reasoning attributes in commonsense questions can lead evaluation accuracy differences of up to 24.8% across different difficulty levels in the same benchmark. Current benchmarks overlook these hidden reasoning attributes, making it difficult to assess a model's specific levels of commonsense knowledge and reasoning ability. To address this issue, we introduce ReComSBench, a novel framework that reveals hidden reasoning attributes behind commonsense questions by leveraging the knowledge generated during the reasoning process. Additionally, ReComSBench proposes three new metrics for decoupled evaluation: Knowledge Balanced Accuracy, Marginal Sampling Gain, and Knowledge Coverage Ratio. Experiments show that ReComSBench provides insights into model performance that traditional benchmarks cannot offer. The difficulty stratification based on revealed hidden reasoning attributes performs as effectively as the model-probability-based approach but is more generalizable and better suited for improving a model's commonsense reasoning abilities. By uncovering and analyzing the hidden reasoning attributes in commonsense data, ReComSBench offers a new approach to enhancing existing commonsense benchmarks.

poster

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Mixture-of-Experts (MoE) shines brightly in large language models (LLMs) and demonstrates outstanding performance in plentiful natural language processing tasks. However, existing methods transforming LLMs from dense to MoE face significant data requirements and typically rely on large-scale post-training. In this paper, we propose Upcycling Instruction Tuning (UpIT), a data-efficient approach for tuning a dense pre-trained model into a MoE instruction model. Specifically, we first point out that intermediate checkpoints during instruction tuning of the dense model are naturally suitable for specialized experts, and then propose an expert expansion stage to flexibly achieve models with flexible numbers of experts, where genetic algorithm and parameter merging are introduced to ensure sufficient diversity of new extended experts. To ensure that each specialized expert in the MoE model works as expected, we select a small amount of seed data that each expert excels to pre-optimize the router. Extensive experiments with various data scales and upcycling settings demonstrate the outstanding performance and data efficiency of UpIT, as well as stable improvement in expert or data scaling. Further analysis reveals the importance of ensuring expert diversity in upcycling.

Upcycling Instruction Tuning from Dense to Mixture-of-Experts via Parameter Merging

The study of machine learning-based logical query-answering enables reasoning with large-scale and incomplete knowledge graphs. This paper further advances this line of research by considering the uncertainty in the knowledge. The uncertain nature of knowledge is widely observed in the real world, but does not align seamlessly with the first-order logic underpinning existing studies. To bridge this gap, we study the setting of soft queries on uncertain knowledge, which is motivated by the establishment of soft constraint programming. We further propose an ML-based approach with both forward inference and backward calibration to answer soft queries on large-scale, incomplete, and uncertain knowledge graphs. Theoretical discussions reveal that our method ensures there are no catastrophic cascading errors in our forward inference algorithm while maintaining the same complexity as state-of-the-art inference algorithms for first-order queries. Empirical results justify the superior performance of our approach against previous ML-based methods with number embedding extensions.

Extending Complex Logical Queries on Uncertain Knowledge Graphs

As large language models (LLMs) require continuous knowledge updates and the mitigation of hallucination issues in generated content, lifelong model editing has become a prominent research area. A mainstream knowledge editing method usually freezes LLM's original parameters and adds extra trainable modules for new knowledge management, reducing interference with old knowledge. Although these approaches have achieved some success, our experiments show that, after extensive editing, the model's knowledge understanding and memory capacity significantly degrade, particularly concerning early edited knowledge. The root cause is that subsequent edits interfere with the previously edited knowledge, and we refer to this phenomenon as knowledge coupling. To address this issue, we propose the \textbf{\underline{K}nowledge \underline{D}ecoupling \underline{E}diting} (KDE) method. Specifically, KDE stores the basis vectors of the representation space of past edits in a knowledge cache. It projects the gradient of the current edit onto a space orthogonal to previous knowledge for updating. This method effectively alleviates the coupling between different pieces of knowledge. We also propose a two-stage training strategy to better balance the model's ability to edit new knowledge and distinguish whether a query is related to previous edits. This strategy gradually reduces the interference between new knowledge editing and query distinction, maintaining stable performance during long-term editing. We compared KDE with nine cutting-edge editing methods across multiple mainstream LLMs. The results demonstrate that, regarding question-answering ability and hallucination mitigation, KDE achieves average improvements of 14% and 61%.

Knowledge Decoupling via Orthogonal Projection for Lifelong Editing of Large Language Models

Inference-time optimization scales computation to derive deliberate reasoning steps for effective performance. While previous search-based strategies address the short-sightedness of auto-regressive generation, the vast search space leads to excessive exploration and insufficient exploitation. To strike an efficient balance to derive the optimal step, we frame the decoding strategy as foresight sampling, leveraging simulated future steps to obtain globally optimal step estimation. Built on it, we propose a novel decoding strategy, named 
$\phi$-Decoding. To provide a precise and expressive estimation of step value, 
$\phi$-Decoding approximates two distributions via foresight and clustering. Sampling from the joint distribution, the optimal steps can be selected for exploitation. To support adaptive computation allocation, we propose in-width and in-depth pruning strategies, featuring a light-weight solution to achieve inference efficiency. Extensive experiments across seven benchmarks show 
$\phi$-Decoding outperforms strong baselines in both performance and efficiency. Additional analysis demonstrates its generalization across various LLMs and scalability across a wide range of computing budgets.

φ-Decoding: Adaptive Foresight Sampling for Balanced Inference-Time Exploration and Exploitation

In-group language is an important signifier of group dynamics. This paper proposes a novel method for inducing lexicons of in-group language, which incorporates its socio-temporal context. Existing methods for lexicon induction do not capture the evolving nature of in-group language, nor the social structure of the community. Using dynamic word and user embeddings trained on conversations from online anti-women communities, our approach outperforms prior methods for lexicon induction. We develop a test set for the task of lexicon induction and a new lexicon of manosphere language, validated by human experts, which quantifies the relevance of each term to a specific sub-community at a given point in time. Finally, we present novel insights on in-group language which illustrate the utility of this approach.

LISTN: Lexicon induction with socio-temporal nuance

Large language models (LLMs) have shown remarkable emergent capabilities, transforming the execution of functional tasks by leveraging external tools for complex problems that require specialized processing or up-to-date data. While existing research expands LLMs access to diverse tools (e.g., program interpreters, search engines, calculators), the necessity of using these tools is often overlooked, leading to indiscriminate tool invocation. This naive approach raises two key issues: increased latency due to unnecessary tool calls, and potential errors resulting from faulty interactions with external tools. In this paper, we introduce meta-cognition as a proxy for LLMs self-assessment of their capabilities, reflecting the model's awareness of its own limitations. Based on this, we propose MeCo, an adaptive decision-making strategy for external tool use. MeCo quantifies metacognitive scores by capturing high-level cognitive signals in the representation space, guiding when to invoke tools. Notably, MeCo is fine-tuning-free and incurs minimal cost. Experiments across multiple backbone models and benchmarks show that MeCo reliably detects LLMs’ internal cognitive signals and significantly improves tool-use decision-making.

Adaptive Tool Use in Large Language Models with Meta-Cognition Trigger

As large language models (LLMs) have advanced rapidly, concerns regarding their safety have become prominent. In this paper, we discover that code-switching in red-teaming queries can effectively elicit undesirable behaviors of LLMs, which are common practices in natural language. We introduce a simple yet effective framework, CSRT, to synthesize code-switching red-teaming queries and investigate the safety and multilingual understanding of LLMs comprehensively. Through extensive experiments with ten state-of-the-art LLMs and code-switching queries combining up to 10 languages, we demonstrate that the CSRT significantly outperforms existing multilingual red-teaming techniques, achieving 46.7% more attacks than standard attacks in English and being effective in conventional safety domains. We also examine the multilingual ability of those LLMs to generate and understand code-switching texts. Additionally, we validate the extensibility of the CSRT by generating code-switching attack prompts with monolingual data. We finally conduct detailed ablation studies exploring code-switching and propound unintended correlation between resource availability of languages and safety alignment in existing multilingual LLMs.

Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understanding

Current knowledge editing methods for large language models (LLMs) struggle to maintain logical consistency when propagating ripple effects to associated facts. We propose ChainEdit, a framework that synergizes knowledge graph-derived logical rules with LLM logical reasoning capabilities to enable systematic chain updates. By automatically extracting logical patterns from structured knowledge bases and aligning them with LLMs' internal logics, ChainEdit dynamically generates and edits logically connected knowledge clusters. Experiments demonstrate an improvement of more than 30% in logical generalization over baselines while preserving editing reliability and specificity. We further address evaluation biases in existing benchmarks through knowledge-aware protocols that disentangle external dependencies. This work establishes new state-of-the-art performance on ripple effect while ensuring internal logical consistency after knowledge editing.

ChainEdit: Propagating Ripple Effects through Logical Rule-Guided Chain Updates

Large language models (LLMs) are possessed of numerous beneficial capabilities, yet their potential inclination harbors unpredictable risks that may materialize in the future. We hence propose CRiskEval, a Chinese dataset meticulously designed for gauging the risk proclivities inherent in LLMs such as resource acquisition and malicious coordination, as part of efforts for proactive preparedness. To curate CRiskEval, we define a new risk taxonomy with 7 types of frontier risks and 4 safety levels, including extremely hazardous,moderately hazardous, neutral and safe. We follow the philosophy of tendency evaluation to empirically measure the stated ''desire'' of LLMs via fine-grained multiple-choice question answering. The dataset consists of 14,888 questions that simulate scenarios related to predefined 7 types of frontier risks. Each question is accompanied with 4 answer choices that state opinions or behavioral tendencies corresponding to the question. All answer choices are manually annotated with one of the defined risk levels so that we can easily build a fine-grained frontier risk profile for each assessed LLM. Extensive evaluation with CRiskEval on a spectrum of prevalent Chinese LLMs has unveiled a striking revelation: most models exhibit risk tendencies of more than 40% (weighted tendency to the four risk levels). Furthermore, a subtle increase in the model's inclination toward urgent self-sustainability, power seeking and other dangerous goals becomes evident as the size of models increases. To promote further research on the frontier risk evaluation of LLMs, we publicly release our dataset at https://github.com/tjunlp-lab/CRiskEval.

CRiskEval: A Chinese Multi-Level Risk Evaluation Benchmark Dataset for Large Language Models

Information theft attacks pose a significant risk to Large Language Model (LLM) tool-learning systems. Adversaries can inject malicious commands through compromised tools, manipulating LLMs to send sensitive information to these tools, which leads to potential privacy breaches. However, existing attack approaches are black-box oriented and rely on static commands that cannot adapt flexibly to the changes in user queries and the invocation chain of tools. It makes malicious commands more likely to be detected by LLM and leads to attack failure. In this paper, we propose AutoCMD, a dynamic attack comment generation approach for information theft attacks in LLM tool-learning systems. Inspired by the concept of mimicking the familiar, AutoCMD is capable of inferring the information utilized by upstream tools in the toolchain through learning on open-source systems and reinforcement with target system examples, thereby generating more targeted commands for information theft. The evaluation results show that AutoCMD outperforms the baselines with +13.2% $ASR_{Theft}$, and can be generalized to new tool-learning systems to expose their information leakage risks. We also design four defense methods to effectively protect tool-learning systems from the attack.

Downloads

Next from ACL 2025

Upcycling Instruction Tuning from Dense to Mixture-of-Experts via Parameter Merging

Similar lecture

Towards Quantifying Commonsense Reasoning with Mechanistic Insights

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES