China

As LLMs excel on standard reading comprehension benchmarks, attention is shifting toward evaluating their capacity for complex abstract reasoning and inference. Literature-based benchmarks, with their rich narrative and moral depth, provide a compelling framework for evaluating such deeper comprehension skills. Here, we present Morables, a human-verified benchmark built from fables and short stories drawn from historical literature. The main task is structured as multiple-choice questions targeting moral inference, with carefully crafted distractors that challenge models to go beyond shallow, extractive question answering. To further stress-test model robustness, we introduce adversarial variants designed to surface vulnerabilities such as data contamination. Our findings show that while larger models outperform smaller ones, they remain susceptible to adversarial manipulation and often rely on superficial patterns rather than true moral reasoning. Interestingly, reasoning-enhanced models fail to bridge this gap, suggesting that scale -- not reasoning ability -- is the primary driver of performance.

EMNLP 2025

Morables: A Benchmark for Assessing Abstract Moral Reasoning in LLMs with Fables

moral inference

fables

llms

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community."<br>

*Message from the General Chair, Dirk Hovy*





<br>

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

poster

Information retrieval (IR) is the task of finding relevant documents in response to a user query. Although Spanish is the second most spoken native language, current IR benchmarks lack Spanish data, which limits the development of information access tools for Spanish speakers. We introduce MessIRve, a large-scale Spanish IR dataset with almost 700 thousand queries from Google's autocomplete API and relevant documents sourced from Wikipedia. MessIRve's queries reflect diverse Spanish-speaking regions, unlike other datasets that are translated from English or do not consider dialectal variations. The large size of the dataset allows it to cover a wide variety of topics, unlike smaller datasets. We provide a comprehensive description of the dataset, comparisons with existing datasets, and baseline evaluations of prominent IR models. Our contributions aim to advance Spanish IR research and improve information access for Spanish speakers.

MessIRve: A Large-Scale Spanish Information Retrieval Dataset

This paper introduces AFRIDOC-MT, a document-level multi-parallel translation dataset covering English and five African languages: Amharic, Hausa, Swahili, Yorùbá, and Zulu. The dataset comprises 334 health and 271 information technology news documents, all human-translated from English to these languages. We conduct document-level translation benchmark experiments by evaluating the ability of neural machine translation (NMT) models and large language models (LLMs) to translate between English and these languages, at both the sentence and pseudo-document levels, the outputs being realigned to form complete documents for evaluation. Our results indicate that NLLB-200 achieves the best average performance among the standard NMT models, while GPT-4o outperforms general-purpose LLMs. Fine-tuning selected models leads to substantial performance gains, but models trained on sentences struggle to generalize effectively to longer documents. Furthermore, our analysis reveals that some LLMs exhibit issues such as under-generation, over-generation, repetition of words and phrases, and off-target translations, specifically for translation into African languages.

AFRIDOC-MT: Document-level MT Corpus for African Languages

Puns are a form of humorous wordplay that exploits polysemy and phonetic similarity. While LLMs have shown promise in detecting puns, we show in this paper that their understanding often remains shallow, lacking the nuanced grasp typical of human interpretation. By systematically analyzing and reformulating existing pun benchmarks, we demonstrate how subtle changes in puns are sufficient to mislead LLMs. Our contributions include comprehensive and nuanced pun detection benchmarks, human evaluation across recent LLMs, and an analysis of the robustness challenges these models face in processing puns.

Pun Unintended: LLMs and the Illusion of Humor Understanding

Large language models (LLMs) demonstrate exceptional performance on tasks requiring complex linguistic abilities, such as reference disambiguation and metaphor recognition/generation. Although LLMs possess impressive capabilities, their internal mechanisms for processing and representing linguistic knowledge remain largely opaque. Prior research on linguistic mechanisms is limited by coarse granularity, limited analysis scale, and narrow focus. In this study, we propose LinguaLens, a systematic and comprehensive framework for analyzing the linguistic mechanisms of large language models, based on Sparse Auto-Encoders (SAEs). We extract a broad set of Chinese and English linguistic features across four dimensions—morphology, syntax, semantics, and pragmatics. By employing counterfactual methods, we construct a large‐scale counterfactual dataset of linguistic features for mechanism analysis. Our findings reveal intrinsic representations of linguistic knowledge in LLMs, uncover patterns of cross‐layer and cross‐lingual distribution, and demonstrate the potential to control model outputs. This work provides a systematic suite of resources and methods for studying linguistic mechanisms, offers strong evidence that LLMs possess genuine linguistic knowledge, and lays the foundation for more interpretable and controllable language modeling in future research.

LinguaLens: Towards Interpreting Linguistic Mechanisms of Large Language Models via Sparse Auto-Encoder

Understanding how Transformer-based language models store and retrieve factual associations is critical for improving interpretability and enabling targeted model editing. Prior work, primarily on GPT-style models, has identified MLP modules in early layers as key contributors to factual recall. However, it remains unclear whether these findings generalize across different autoregressive architectures. To address this, we conduct a comprehensive evaluation of factual recall across several models---including GPT, LLaMA, Qwen, and DeepSeek---analyzing where and how factual information is encoded and accessed. Consequently, we find that Qwen-based models behave differently from previous patterns: attention modules in the earliest layers contribute more to factual recall than MLP modules. Our findings suggest that even within the autoregressive Transformer family, architectural variations can lead to fundamentally different mechanisms of factual recall.

Do All Autoregressive Transformers Remember Facts the Same Way? A Cross-Architecture Analysis of Recall Mechanisms

Pinpointing the subset of general data from large corpora that optimally enhances specific capabilities in large language models (LLMs) offers an effective solution to address domain-specific data scarcity and reduce training costs. However, mainstream methods that rely on observing changes in benchmark metrics or gradient similarity in the training phase are based on superficial observations and fail to link LLMs capabilities with their internal components, thus lacking interpretability, resulting in inefficient and limited performance gains. In this work, we propose a lightweight and high interpretability method to associate LLMs capabilities with internal components, specifically identifying a correspondence between specific capabilities and attention heads. We first delineate five fundamental application capabilities of LLMs. Probing techniques are then employed to identify the specific attention heads corresponding to each, thereby establishing a mapping between these capabilities and internal model components. For targeted instruction tuning, we subsequently decompose the complex abilities required for intricate tasks into combinations of these fundamental capabilities and directly select data using corresponding attention heads. Experiments on LLaMA3.1-8B and Qwen2.5-7B achieve a discrimination accuracy of over 70% for various capabilities. And results on the MMLU and BBH datasets show that our method outperforms the gradient-based selection method LESS by 1.3%-2% and other intermediate-state-based selection methods by 5%-6%.

Probing and Boosting Large Language Models Capabilities via Attention Heads

N-ary Knowledge Graphs (NKGs) are a specialized type of knowledge graph designed to efficiently represent complex real-world facts. Unlike traditional knowledge graphs, where a fact typically involves two entities, NKGs can capture n-ary facts containing more than two entities. Link prediction in NKGs aims to predict missing elements within these n-ary facts, which is essential for completing NKGs and improving the performance of downstream applications. This task has recently gained significant attention. In this paper, we present the first comprehensive survey of link prediction in NKGs, providing an overview of the field, systematically categorizing existing methods, and analyzing their performance and application scenarios. We also outline promising directions for future research.

A Survey of Link Prediction in N-ary Knowledge Graphs

Braille plays a vital role in education and information accessibility for visually impaired individuals. However, Braille information processing faces challenges such as data scarcity and ambiguities in mixed-text contexts. We construct English and Chinese Braille Mixed Datasets (EBMD/CBMD) with mathematical formulas to support diverse Braille domain research, and propose a syntax tree-based augmentation method tailored for Braille data. To address the underperformance of traditional fine-tuning methods in braille-related tasks, we investigate Braille Knowledge-Based Fine-Tuning (BKFT), which reduces the learning difficulty of Braille contextual features. BrailleLLM employs BKFT via instruction tuning to achieve unified Braille translation, formula-to-Braille conversion, and mixed-text translation. Experiments demonstrate that BKFT achieves significant performance improvements over conventional fine-tuning in Braille translation scenarios. Our open-sourced datasets and methodologies establish a foundation for low-resource multilingual Braille research.

BrailleLLM: Braille Instruction Tuning with Large Language Models for Braille Domain Tasks

Providing constructive feedback to paper authors is a core component of peer review. With reviewers increasingly having less time to perform reviews, automated support systems are required to ensure high reviewing quality, thus making the feedback in reviews useful for authors. To this end, we identify four key aspects of review comments (individual points in weakness sections of reviews) that drive the utility for authors: Actionability, Grounding & Specificity, Verifiability, and Helpfulness. To enable evaluation and development of models assessing review comments, we introduce the RevUtil dataset. We collect 1,430 human-labeled review comments and scale our data with 10k synthetically labeled comments for training purposes. The synthetic data additionally contains rationales, i.e., explanations for the aspect score of a review comment. Employing the RevUtil dataset, we benchmark fine-tuned models for assessing review comments on these aspects and generating rationales. Our experiments demonstrate that these fine-tuned models achieve agreement levels with humans comparable to, and in some cases exceeding, those of powerful closed models like GPT-4o. Our analysis further reveals that GPT-4-generated reviews generally underperform human reviews on our four aspects.

The Good, the Bad and the Constructive: Automatically Measuring Peer Review's Utility for Authors

Large Language Models (LLMs) require robust evaluation. However, existing frameworks often rely on curated datasets that, once public, may be accessed by newer LLMs. This creates a risk of data leakage, where test sets inadvertently become part of training data, compromising evaluation fairness and integrity. To mitigate this issue, we propose Behave as Claimed (BaC), a novel evaluation framework inspired by counterfactual reasoning. BaC constructs a ``what-if'' scenario where LLMs respond to counterfactual questions about how they would behave if the input were manipulated. We refer to these responses as claims, which are verifiable by observing the LLMs' actual behavior when given the manipulated input. BaC dynamically generates and verifies counterfactual questions using various few-shot in-context learning evaluation datasets, reducing their susceptibility to data leakage. Moreover, BaC provides a more challenging evaluation paradigm for LLMs. LLMs must thoroughly understand the prompt, the task, and the consequences of their responses to achieve better performance. We evaluate several state-of-the-art LLMs and find that, while most perform well on the original datasets, they struggle with BaC. This suggests that LLMs usually fail to align their claims with their actual behavior and that high performance on standard datasets may be less stable than previously assumed.

Downloads

Next from EMNLP 2025

MessIRve: A Large-Scale Spanish Information Retrieval Dataset

Similar lecture

Structured Moral Reasoning in Language Models: A Value-Grounded Evaluation Framework