Austria

Recent studies on semantic frame induction have demonstrated that the emergence of pre-trained language models (PLMs) has led to more accurate results.
However, most existing studies evaluate the performance using frame resources such as FrameNet, which may not accurately reflect real-world language usage.
In this study, we conduct semantic frame induction using the Colossal Clean Crawled Corpus (C4) and assess the applicability of existing frame induction methods to real-world data.
Our experimental results demonstrate that existing frame induction methods are effective on real-world data and that frames corresponding to novel concepts can be induced.

ACL 2025

Semantic Frame Induction from a Real-World Corpus

frame induction

frame semantics

framenet

## 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." 

*Message from the General Chair, Dirk Hovy*

EMNLP 2025

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

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)

Morphological defectivity is an intriguing and understudied phenomenon in linguistics. Addressing defectivity, where expected inflectional forms are absent, is essential for improving the accuracy of NLP tools in morphologically rich languages. However, traditional linguistic resources often lack coverage of morphological gaps as such knowledge requires significant human expertise and effort to document and verify. For scarce linguistic phenomena in under-explored languages, Wikipedia and Wiktionary often serve as among the few accessible resources. Despite their extensive reach, their reliability has been a subject of controversy. This study customizes a novel neural morphological analyzer to annotate Latin and Italian corpora. Using the massive annotated data, crowd-sourced lists of defective verbs compiled from Wiktionary are validated computationally. Our results indicate that while Wiktionary provides a highly reliable account of Italian morphological gaps, 7% of Latin lemmata listed as defective show strong corpus evidence of being non-defective. This discrepancy highlights potential limitations of crowd-sourced wikis as definitive sources of linguistic knowledge, particularly for less-studied phenomena and languages, despite their value as resources for rare linguistic features. By providing scalable tools and methods for quality assurance of crowd-sourced data, this work advances computational morphology and expands linguistic knowledge of defectivity in non-English, morphologically rich languages.

Lost and Found: Computational Quality Assurance of Crowdsourced Knowledge on Morphological Defectivity in Wiktionary

Large Language Models (LLMs) exhibit strong conversational abilities but often generate falsehoods. Prior work suggests that the truthfulness of simple propositions can be represented as a single linear direction in a model's internal activations, but this may not fully capture its underlying geometry. In this work, we extend the concept cone framework—recently introduced for modeling refusal—to the domain of truth. We identify multi-dimensional cones that causally mediate truth-related behavior across multiple LLM families. Our results are supported by three lines of evidence: (i) causal interventions reliably flip model responses to factual statements; (ii) learned cones generalize across model architectures; and (iii) cone-based interventions preserve unrelated model behavior. These findings reveal the richer, multidirectional structure governing simple true/false propositions in LLMs and highlight concept cones as a promising tool for probing abstract behaviors.

From Directions to Cones: Multidimensional Representations of Propositional Facts in LLMs

Congenital heart disease (CHD) presents complex, lifelong challenges often underrepresented in traditional clinical metrics. While unstructured narratives offer rich insights into patient and caregiver experiences, manual thematic analysis (TA) remains labor-intensive and unscalable. We propose a fully automated large language model (LLM) pipeline that performs end-to-end TA on clinical narratives which eliminates the need for manual coding or full transcript review. Our system employs a novel multi-agent framework, where specialized LLM agents assume roles to enhance theme quality and alignment with human analysis. To further improve thematic relevance, we optionally integrate reinforcement learning from human feedback (RLHF). This supports scalable, patient-centered analysis of large qualitative datasets and allows LLMs to be fine-tuned for specific clinical contexts.

Auto-TA: Towards Scalable Automated Thematic Analysis (TA) via Multi-Agent Large Language Models with Reinforcement Learning

Deterministically controlling the target generation language of large multilingual language models (LLMs) remains a fundamental challenge, particularly in zero-shot settings where neither explicit language prompts nor fine-tuning are available. In this work, we investigate whether sparse autoencoder (SAE) features, previously shown to correlate with interpretable model behaviors, can be leveraged to steer the generated language of LLMs during inference. Leveraging pretrained SAEs on the residual streams of Gemma-2B and Gemma-9B, we identify features whose activations differ most significantly between English and four target languages: Chinese, Japanese, Spanish, and French. By modifying just a single SAE feature at one transformer layer, we achieve controlled language shifts with up to 90% success, as measured by FastText language classification, while preserving semantic fidelity according to LaBSE (Language-Agnostic BERT Sentence Embedding) similarity. Our analysis reveals that language steering is most effective in mid-to-late transformer layers and is amplified by specific attention heads disproportionately associated with language-sensitive SAE features. These results demonstrate the promise of sparse feature steering as a lightweight and interpretable mechanism for controllable multilingual generation.

Causal Language Control in Multilingual Transformers via Sparse Feature Steering

Variation in human perspectives has drawn increasing attention in natural language processing. The widespread human annotation disagreement challenges the traditional paradigm of a single "ground truth" and raises concerns about the limitations of conventional label aggregation methods and the uniform models built upon them, which often discard minority opinions and obscure valuable individual perspectives. This thesis proposal investigates three core dimensions of perspective-oriented research: (1) annotation formats that better capture the nuance and uncertainty of individual judgments; (2) modeling approaches that leverage socio-demographic features to improve prediction for underrepresented or minority viewpoints; and (3) personalized generation that tailor outputs to individual users’ preferences and communicative styles. Through this work, we aim to advance methods that more faithfully reflect the diversity of human interpretation, enhancing both inclusiveness and fairness in language technologies.

Proposal: From One-Fit-All to Perspective Aware Modeling

We investigate feature universality in Gemma-2 language models (Gemma-2-2B & Gemma-2-9B), asking whether models with a fourfold difference in scale still converge on comparable internal concepts. Using the sparse autoencoder (SAE) dictionary learning pipeline, we used pretrained SAEs on each model’s residual-stream activations, aligned the resulting monosemantic features via activation correlation, and compared the matched feature spaces with metrics such as SVCCA and RSA. Middle layers yield the strongest overlap, indicating that this is where both models most similarly represent concepts, while early and late layers show much less similarity. Preliminary experiments extending the analysis from single tokens to multi-token subspaces show that semantically similar subspaces tend to interact similarly with LLMs. These results offer further evidence that large language models carve the world into broadly similar, interpretable features despite size differences, reinforcing universality as a foundation for cross-model interpretability.

Semantic Convergence: Investigating Shared Representations Across Scaled LLMs

Large language models often suffer from language confusion, a phenomenon where responses are partially or entirely generated in unintended languages. This can critically impact user experience in low-resource settings. We hypothesize that conventional supervised fine-tuning exacerbates this issue because the softmax objective focuses probability mass only on the single correct token but does not explicitly penalize cross-lingual mixing. 
Interestingly, by observing loss trajectories during the pretraining phase, we observe that models fail to learn to distinguish between monolingual and language-confused text. Additionally, we find that ORPO, which adds penalties for unwanted output styles to standard SFT, effectively suppresses language-confused generations even at high decoding temperatures without degrading overall model performance. Our findings suggest that incorporating appropriate penalty terms can mitigate language confusion in low-resource settings with limited data.

Controlling Language Confusion in Multilingual LLMs

We introduce a modified sequence tagging architecture, proposed in (Omelianchuk et al., 2020), for the Grammatical Error Correction of the Russian language. We propose language-specific operation set and preprocessing algorithm as well as a classification scheme which makes distinct predictions for insertions and other operations. The best versions of our models outperform previous approaches and set new SOTA on the two Russian GEC benchmarks -- RU-Lang8 and GERA, while achieve competitive performance on RULEC-GEC.

Grammatical Error Correction via Sequence Tagging for Russian

Text corpora in non-English clinical contexts are sparse, where synthetic data generation with Large Language Models (LLMs) appears as a promising strategy to overcome this data gap. In order to test the quality of LLMs in generating synthetic data, we applied them to our novel German Medical Interview Questions Corpus (GerMedIQ), consisting of 4,524 unique question-response pairs in German. We augmented our corpus by asking a cohort of models to produce suitable responses to the same questions. Structural and semantic evaluations of the synthetic responses revealed that while augmented responses may meet the grammatical requirements, most models were not able to produce semantically comparable responses to humans. Also, an LLM-as-a-judge experiment showcased that human responses were consistently rated more appropriate than synthetic ones. We find that data augmentation with LLMs in non-English and clinical domain contexts has to be performed carefully.

GerMedIQ: A Resource for Simulated and Synthesized Anamnesis Interview Responses in German

This paper examines differences between stream-of-consciousness narratives written by humans and those generated by large language models (LLMs) to assess narrative coherence and personality expression. We generated texts by prompting LLMs (Llama-3.1-8B \& DeepSeek-R1-Distill-Llama-8B) with the first half of essays either revealing the personality characteristics (Big Five) or hiding them. Our analysis revealed consistently low similarity between LLM-generated continuations and original human texts, as measured by cosine similarity, perplexity, and BLEU scores. Including explicit personality traits significantly enhanced Llama-3.1-8B's performance, particularly in BLEU scores. Further analysis of personality expression showed varying alignment patterns between LLMs and human texts. Further analysis of personality expression found higher openness for all models. Additionally, Llama-3.1-8B models exhibited higher Extraversion but low Agreeableness, while DeepSeek-R1-Distill-Llama-8B's showed dramatic personality shifts during its thinking process, particularly when provided with personality traits in the prompt.

Downloads

Next from ACL 2025

Lost and Found: Computational Quality Assurance of Crowdsourced Knowledge on Morphological Defectivity in Wiktionary

Similar lecture

FrameEOL: Semantic Frame Induction using Causal Language Models