Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 2: Short Papers)

When Does Auxiliary Modality Matter in Solving Geometric Problems? A Comprehensive Study of Textual, Formal, and Visual Modalities (2026.eacl-short)

Challenge:	Recent large language models claim long context windows, but evaluations often involve simple retrieval tasks or synthetic tasks padded with irrelevant text.
Approach:	They use grammars to generate simplified English with logical representations to create long input text while controlling its semantics.
Outcome:	The proposed model performs better with realistic distractors than with standard models.

Copied to clipboard

Hyuk Namgoong, Jeesu Jung, Yerim Han, Sangkeun Jung

Progressive Visual Refinement for Multi-modal Summarization (2026.eacl-short)

Challenge:	Large Language Models (LLMs) face challenges in integrating linguistic and spatial reasoning, which limits their performance on geometry problems.
Approach:	They compare four auxiliary modalities on open- and closed-source multimodal LLMs . they show that DES boosts the accuracy of open-source LLM models .
Outcome:	The proposed modalities improve performance on open- and closed-source LLMs.

Copied to clipboard

Ye Xiong, Hidetaka Kamigaito, Soichiro Murakami, Peinan Zhang, Hiroya Takamura, Manabu Okumura

Benchmarking Temporal Reasoning and Alignment Across Chinese Dynasties (2026.eacl-short)

Challenge:	Multi-modal summarization (MMS) is a critical research area driven by the proliferation of multimedia content.
Approach:	They propose a patch-refined visual information network to exploit multimodal information . they propose combining visual information with textual information to generate concise summaries .
Outcome:	Extensive experiments on two public MMS datasets show the superiority of the proposed model.

Copied to clipboard

Zhenglin Wang, Jialong Wu, Pengfei Li, Yong Jiang, Deyu Zhou

Training in Step-by-Step Formal Reasoning Improves Pronominal Reasoning in Language Models (2026.eacl-short)

Challenge:	Existing temporal reasoning benchmarks rely on rule-based construction and lack contextual depth . a recent study found existing LLMs struggle with nuanced temporal understanding .
Approach:	a benchmark is designed to evaluate LLMs on temporal reasoning in Chinese dynasties.
Outcome:	a new benchmark evaluates LLMs on temporal reasoning across Chinese dynasties . it emphasizes cross-entity relationships, pairwise temporal alignment, contextualized and culturally-grounded reasoning . results show existing LLM benchmarks struggle with nuanced temporal understanding .

Copied to clipboard

Vagrant Gautam

When Words Wear Masks: Detecting Malicious Intents and Hostile Impacts of Online Hate Speech (2026.eacl-short)

Challenge:	Large reasoning models are limited to formal reasoning, i.e., math, code, and logic.
Approach:	They evaluate a set of large reasoning models on a dataset for pronoun resolution and fidelity.
Outcome:	The results show that distilling step-by-step formal reasoning improves pronoun resolution and fidelity.

Copied to clipboard

Priyansh Singhal, Piyush Joshi

Lost in Activations: A Neuron-level Analysis of Encoders for Cross-Lingual Emotion Detection (2026.eacl-short)

Challenge:	Existing methods for hate speech detection treat hate speech as a monolithic phenomenon, ignoring the speaker’s motivations and potential societal consequences.
Approach:	They propose a dataset with a dual taxonomy that separates Intent (why the speaker produced hate speech) and Impact (what harm it may cause to individuals and communities) they propose to use this data to enable content moderation and user safety.
Outcome:	The proposed dataset captures Intent (why the speaker produced hate speech) and Impact (what harm it may cause to individuals and communities) of online hateful posts.

Copied to clipboard

Pranaydeep Singh, Orphee De Clercq, Els Lefever

Multilinguality

Challenge:	XLM-R models for multilingual emotion classification are still lacking in understanding of their internal decision-making mechanisms.
Approach:	They propose to use neuron-level activation analysis to study the inter-language differences between neurons.
Outcome:	The proposed model consistently encodes emotion-related concepts across languages, but others show strong monolingual specialization.

McMining: Automated Discovery of Misconceptions in Student Code (2026.eacl-short)

Copied to clipboard

Erfan Al-Hossami, Razvan Bunescu

Surprisal from Larger Transformer-based Language Models Predicts fMRI Data More Poorly (2026.eacl-short)

Challenge:	misconceptions can lead to bugs and slow down learning of related concepts . a misconception is a belief in a false statement, such as believing that the earth is flat or that real numbers are countable.
Approach:	They propose a task of mining programming misconceptions from student code samples . they introduce McMining, which uses a benchmark dataset to identify misconceptions .
Outcome:	The proposed models are effective at finding misconceptions in student code.

Copied to clipboard

Yi-Chien Lin, William Schuler

WebRollback: Enhancing Web Agents with Explicit Rollback Mechanisms (2026.eacl-short)

Challenge:	Recent work has observed an inverse scaling relationship between Transformers’ per-word estimated probability and the predictive power of their surprisal estimates on reading times.
Approach:	They conducted a more comprehensive evaluation using surprisal estimates from 17 pre-trained LMs on two functional magnetic resonance imaging datasets.
Outcome:	Recent work shows that surprisal from larger Transformer-based models is less predictive of reading times, resolving the inconclusive results and indicating that this trend is not specific to latency-based measures.

Copied to clipboard

Zhisong Zhang, Tianqing Fang, Kaixin Ma, Wenhao Yu, Hongming Zhang, Haitao Mi, Dong Yu

Challenge:	Recent studies have adopted a greedy one-way search strategy to deal with dynamic web environments.
Approach:	They propose to integrate a rollback mechanism into web agents to allow them to revert back to a previous state in navigation trajectory.
Outcome:	The proposed method is able to revert back to a previous state in its navigation trajectory, allowing the models to directly control the search process.

Hacking Neural Evaluation Metrics with Single Hub Text (2026.eacl-short)

Copied to clipboard

Hiroyuki Deguchi, Katsuki Chousa, Yusuke Sakai

Machine Translation

Challenge:	Recent embedding-based neural text evaluation metrics are not reliable due to black-box nature of neural networks.
Approach:	They propose to find a single adversarial text in the discrete space that is consistently evaluated as high-quality regardless of the test cases.
Outcome:	The proposed method outperforms translations generated individually for each source sentence in English-to-Japanese and English- to-German translation tasks.

To Paraphrase or Not: Efficient Comment Detoxification with Unsupervised Detoxifiability Discrimination (2026.eacl-short)

Copied to clipboard

Jing Ke, Zheyong Xie, Shaosheng Cao, Tong Xu, Enhong Chen

Hey, wait a minute: on at-issue sensitivity in Language Models (2026.eacl-short)

Challenge:	Existing methods for detoxification of toxic comments are limited by overcorrection and data scarcity . experimental results show that DID outperforms existing methods on academic data and an industrial platform .
Approach:	They propose a paradigm that adaptively conducts filtering or paraphrasing for each toxic comment based on its detoxifiability . they propose 'detoxifiabilities-aware detoxification' that can be trained to filter or paraphrase toxic comments based upon their detoxifikatability based only on detoxificable comments .
Outcome:	Experimental results show that DID outperforms existing methods on academic and industrial data.

Copied to clipboard

Sanghee J. Kim, Kanishka Misra

Dialogue Systems & Conversational Agents

Challenge:	Existing methods to evaluate dialogue naturalness are limited.
Approach:	They propose a method to assess dialogue naturalness using linguistic notion of at-issueness.
Outcome:	The proposed method mitigates bias in linguistic analyses of LMs and tests discourse-sensitive behavior.

Exploring Cross-Lingual Voice Conversion Methods for Anonymizing Low-Resource Text-to-Speech (2026.eacl-short)

Copied to clipboard

Shenran Wang, Aidan Pine, Mengzhe Geng

Korean Canonical Legal Benchmark: Toward Knowledge-Independent Evaluation of LLMs’ Legal Reasoning Capabilities (2026.eacl-short)

Challenge:	a growing number of speech synthesis systems clone a person's voice, a new study finds . a variety of voice conversion techniques can mask speaker identities in low-resource text-to-speech systems.
Approach:	They compare voice conversion techniques to mask speaker identities in text-to-speech systems . they build and evaluate speaker-anonymized systems for two Canadian Indigenous languages .
Outcome:	The proposed methods are compared with other approaches for using voice conversion to mask speaker identities in low-resource text-to-speech systems.

Copied to clipboard

Hongseok Oh, Wonseok Hwang, Kyoung-Woon On

Task-Level Instructions Induction for Audio Question Answering from Few Examples (2026.eacl-short)

Challenge:	Large reasoning models trained to reason explicitly in the verbal space have shown superior performance over general large language models (Guo et al., 2025).
Approach:	They propose to use Korean Canonical Legal Benchmark to assess language models' legal reasoning capabilities independently of domain-specific knowledge.
Outcome:	The proposed benchmark outperforms general-purpose models in a systematic evaluation of 30+ models.

Copied to clipboard

Po-Chun Chen, Hen-Hsen Huang, Hsin-Hsi Chen

Optical Character Recognition for the International Phonetic Alphabet (2026.eacl-short)

Challenge:	Large audio-language models benefit from Chain-of-Thought (CoT) prompting for audio question answering (AQA) however, acquiring audio CoT examples is difficult as it requires sequential listening and careful integration of acoustic and linguistic information.
Approach:	They propose a method which induces reusable task instructions from few audio examples once per task.
Outcome:	Evaluated on 9 LALMs across two benchmarks, Audio-Induct outperforms state-of-the-art prompting methods while maintaining low inference costs.

Copied to clipboard

Shu Okabe, Dejvi Zelo, Alexander Fraser

Syntactic Text Processing

Challenge:	Grammar books are increasingly used as additional reference resources for low-resource languages . a significant portion of these documents come from scans and require an OCR tool .
Approach:	They compare two neural OCR frameworks and a large vision-language model with a synthetic dataset based on Wiktionary to study the International Phonetic Alphabet (IPA).
Outcome:	The proposed model improves on the International Phonetic Alphabet (IPA) character set.

How DDAIR you? Disambiguated Data Augmentation for Intent Recognition (2026.eacl-short)

Copied to clipboard

Galo Castillo-López, Alexis Lombard, Nasredine Semmar, Gaël de Chalendar

Information Retrieval

Challenge:	Large Language Models (LLMs) produce ambiguous examples with regard to untargeted classes.
Approach:	They propose to use a sentence transformer to detect ambiguous augmented examples generated by Large Language Models for intent recognition.
Outcome:	The proposed method improves the quality of augmented data generated by large language models in low-resource scenarios.

Measuring Linguistic Competence of LLMs on Indigenous Languages of the Americas (2026.eacl-short)

Copied to clipboard

Justin Vasselli, Arturo Martínez Peguero, Frederikus Hudi, Haruki Sakajo, Taro Watanabe

Challenge:	Existing benchmarks for linguistic knowledge of Indigenous languages of the Americas focus on high- and medium-resource languages with substantial digital presence.
Approach:	They propose a framework for probing large language models’ linguistic knowledge of Indigenous languages of the Americas using zero-shot prompting and few-shot probing.
Outcome:	The proposed framework evaluates models from five major families on 13 Indigenous languages including Bribri, Guarani, and Nahuatl.

Morpheme Matters: Morpheme-Based Subword Tokenization for Korean Language Models (2026.eacl-short)

Copied to clipboard

DongHyeok Lee, Jeongyeon Park, Kyungbeen Cho, Jae Sung Lee

Syntactic Text Processing

Challenge:	Existing tokenizers rely on frequency-based segmentation to represent words . this often leads to inefficient token representations and oversegmentation .
Approach:	They propose a tokenization method that emphasizes the importance of Korean morphological structures in eojeol.
Outcome:	The proposed method outperforms existing tokenizers on Korean benchmark tasks and produces significantly fewer tokens per input sequence.

Communication Enables Cooperation in LLM Agents: A Comparison with Curriculum-Based Approaches (2026.eacl-short)

Copied to clipboard

Hachem Madmoun, Salem Lahlou

Mind Your Special Tokens! On the Importance of Dedicated Sequence-End Tokens in Vision-Language Embedding Models (2026.eacl-short)

Challenge:	a "cheap talk" channel increases cooperation in 4-player Stag Hunt, but a complex curriculum can induce "learned pessimism" in agents.
Approach:	They investigate whether a direct communication channel can elicit cooperation in multi-agent LLMs.
Outcome:	The proposed curriculum reduces agent payoffs by 27.4% in a 4-player Stag Hunt simulation.

Copied to clipboard

Elio Musacchio, Giovanni Semeraro, Goran Glavaš

The Reasoning Lingua Franca: A Double-Edged Sword for Multilingual AI (2026.eacl-short)

Challenge:	Large Vision-Language Models (LVLMs) are highly sensitive to end-of-input artifacts in fine-tuning and inference data, e.g., whether input sequences end with punctuation or newline characters.
Approach:	They propose to convert generative LVLMs into vision-language encoders via contrastive learning objectives and use supervised contrastive objectives to train them.
Outcome:	The proposed approach improves visual and text representations and improves retrieval and (semantic) similarity tasks.

Copied to clipboard

Alan Saji, Raj Dabre, Anoop Kunchukuttan, Ratish Puduppully

Multilinguality

Challenge:	Large Reasoning Models (LRMs) are highly effective on mathematical, scientific, and other question-answering tasks.
Approach:	They compare an LRM's reasoning in English to that of a multilingual question . they find that English reasoning traces exhibit a substantially higher presence of cognitive behaviors .
Outcome:	The LRMs generate reasoning sequences in English, but the language of the question is not.

When Flores Bloomz Wrong: Cross-Direction Contamination in Machine Translation Evaluation (2026.eacl-short)

Copied to clipboard

David Tan, Pinzhen Chen, Josef Van Genabith, Koel Dutta Chowdhury

Multilinguality

Challenge:	Large language models (LLMs) can be benchmark-contaminated, resulting in inflated scores that mask memorization as generalization.
Approach:	They use the FLORES-200 translation benchmark as a diagnostic to investigate cross-direction data contamination.
Outcome:	The proposed model can be cross-directional, boosting performance in unseen translation directions due to target-side memorization.

SAFARI: A Community-Engaged Approach and Dataset of Stereotype Resources in the Sub-Saharan African Context (2026.eacl-short)

Copied to clipboard

Aishwarya Verma, Laud Ammah, Olivia Nercy Ndlovu Lucas, Andrew Zaldivar, Vinodkumar Prabhakaran, Sunipa Dev

STREAM-ZH: Simplified Topic Retrieval Exploration and Analysis Module for Chinese Language (2026.eacl-short)

Challenge:	Existing data collection approaches to generative AI are inadequate to assess its safety and utility.
Approach:	They propose a multilingual stereotype resource that uses socioculturally-situated, community-engaged methods to assess the region’s linguistic diversity and traditional orality.
Outcome:	The proposed method covers four sub-Saharan African countries that are severely underrepresented in NLP resources: Ghana, Kenya, Nigeria, and South Africa.

Copied to clipboard

Hongyi Li, Jianjun Lian, Anton Frederik Thielmann, Andre Python

Information Extraction & Text Mining

Challenge:	Simplified Topic Retrieval Exploration and Analysis Module for Chinese language is the first topic modeling package to fully support the Chinese language.
Approach:	They propose a topic modeling package that fully supports the Chinese language . they use preprocessed textual datasets to assess topic models .
Outcome:	The proposed framework outperforms existing topic models using English-translated textual input.

MiSCHiEF: A Benchmark in Minimal-Pairs of Safety and Culture for Holistic Evaluation of Fine-Grained Image-Caption Alignment (2026.eacl-short)

Copied to clipboard

Sagarika Banerjee, Tangatar Madi, Advait Swaminathan, Jolie Nguyen, Shivank Garg, Kevin Zhu, Vasu Sharma

Exploring Generative Process Reward Modeling for Semi-Structured Data: A Case Study of Table Question Answering (2026.eacl-short)

Challenge:	Fine-grained image-caption alignment is crucial for vision-language models in socially critical contexts.
Approach:	They present a benchmarking dataset for fine-grained image-caption alignment in safety and culture contexts.
Outcome:	The proposed benchmarks show that models perform better at confirming correct pairs than rejecting incorrect ones on dual alignment tasks.

Copied to clipboard

Lei Tang, Wei Zhou, Mohsen Mesgar

Out of Distribution, Out of Luck: Process Rewards Misguide Reasoning Models (2026.eacl-short)

Challenge:	Recent advances in process reward models (PRMs) have demonstrated remarkable improvements in the reasoning capabilities of large language models (LLMs).
Approach:	They evaluate state-of-the-art generative PRMs on table question answering from answer and step perspectives and compare their results to previous studies.
Outcome:	The proposed model can aid solution selection but struggle to generalize to out-of-domain data.

Copied to clipboard

Alexey Dontsov, Anton Korznikov, Andrey V. Galichin, Elena Tutubalina

Learning to Ideate for Machine Learning Engineering Agents (2026.eacl-short)

Challenge:	80% of reasoning model outputs respond to formatting artifacts rather than mathematical content.
Approach:	They evaluate process reward models that provide step-level feedback during inference . they identify distinct reward prediction patterns that differentiate reasoning from non-reasoning model outputs .
Outcome:	The proposed model fails to enhance and sometimes degrade reasoning model performance.

Copied to clipboard

Yunxiang Zhang, Kang Zhou, Zhichao Xu, Kiran Ramnath, Yun Zhou, Sangmin Woo, Haibo Ding, Lin Lee Cheong

Challenge:	Existing machine learning engineering (MLE) agents struggle to iteratively optimize their implemented algorithms for effectiveness.
Approach:	They propose a framework that separates ideation from implementation that allows an implementation agent to request strategic help from a dedicated Ideator.
Outcome:	The proposed framework outperforms implementation-only agent baselines on MLE-Bench and can be trained with reinforcement learning to generate more effective ideas.

Infinity-MoE: Generalizing Mixture of Experts to Infinite Experts (2026.eacl-short)

Copied to clipboard

Shota Takashiro, Takeshi Kojima, Shohei Taniguchi, Yusuke Iwasawa, Yutaka Matsuo

Beyond Tokens: Concept-Level Training Objectives for LLMs (2026.eacl-short)

Challenge:	Existing methods to increase the number of experts are -MoE and .
Approach:	They propose a mixture of experts that selects a few feed-forward networks per token to increase the number of experts.
Outcome:	The proposed model improves on a GPT-2 Small model with 129M active and 186M total parameters by 2.5% over the current model.

Copied to clipboard

Laya Iyer, Pranav Somani, Alice Guo, Dan Jurafsky, Chen Shani

Challenge:	Large language models (LLMs) are trained with a surprisingly narrow objective: predicting the next token in a sequence.
Approach:	They propose a shift from token-level to concept-level prediction where concepts group multiple surface forms of the same idea.
Outcome:	The proposed model improves on human-level models on diverse NLP benchmarks.

Persuasion Tokens for Editing Factual Knowledge in LLMs (2026.eacl-short)

Copied to clipboard

Paul Youssef, Christin Seifert, Jörg Schlötterer

Language Family Matters: Evaluating SpeechLLMs Across Linguistic Boundaries (2026.eacl-short)

Challenge:	In-context knowledge editing (IKE) relies on fact-specific demonstrations which consume significant context window space.
Approach:	They introduce persuasion tokens (P-Tokens) which replicate the effect of IKE demonstrations and allow efficient knowledge editing without requiring fact-specific demonstrations.
Outcome:	The proposed tokens perform comparable to and often exceed IKE on two editing datasets and three LLMs and increase the number of tokens increases performance.

Copied to clipboard

Yuchen Zhang, Ravi Shekhar, Haralambos Mouratidis

When Benchmarks Age: Temporal Misalignment through Large Language Model Factuality Evaluation (2026.eacl-short)

Challenge:	Existing approaches to integrate speech encoders with large language models (LLMs) have limited resources and lack linguistic relatedness.
Approach:	They propose a connector-sharing strategy based on linguistic family membership that allows one connector per family to share a frozen speech encoder with a pretrained LLM.
Outcome:	The proposed system reduces parameter count while improving generalization across domains, compared with existing connectors.

Copied to clipboard

Xunyi Jiang, Dingyi Chang, Julian McAuley, Xin Xu

On the Additive Compositionality of Task Vectors in Vision–Language Models (2026.eacl-short)

Challenge:	Existing studies on LLM factuality evaluation have not investigated the reliability of static evaluation benchmarks.
Approach:	They examine five popular factuality benchmarks and eight LLMs released over different years to assess their reliability.
Outcome:	The proposed method compared five popular factuality benchmarks and eight LLMs released over different years.

Copied to clipboard

Yuting Shi, Houjing Wei, Naoya Inoue

Funny or Persuasive, but Not Both: Evaluating Fine-Grained Multi-Concept Control in LLMs (2026.eacl-short)

Challenge:	In-context learning (ICL) in large language models (LLMs) has been shown to operate through task vectors, but its extension to vision-language models (VLMs) remains underexplored.
Approach:	They construct visual reasoning tasks with clearly defined subtasks and extract task vectors from few-shot demonstrations.
Outcome:	The proposed model can be extended to vision-language models (VLMs) by adding the vectors of its constituent subtasks.

Copied to clipboard

Arya Labroo, Ivaxi Sheth, Vyas Raina, Amaani Ahmed, Mario Fritz

Common Sense or Ableism? Rethinking Commonsense Reasoning Through the Lens of Disability (2026.eacl-short)

Challenge:	Large Language Models (LLMs) provide strong generative capabilities, but many applications require explicit and fine-grained control over specific textual concepts.
Approach:	They propose a framework for fine-grained controllability for single- and dual-concept scenarios . they find performance drops in the dual-constituency setting, even though chosen concepts should be separable .
Outcome:	The proposed framework shows that models struggle with compositionality even when concepts are intuitively independent.

Copied to clipboard

Karina H Halevy, Kimi Wenzel, Seyun Kim, Kyle Dean Bauer, Bruno Neira, Mona T. Diab, Maarten Sap

Machine translation Evaluation Eng-Thai MQM Ranking dataset (2026.eacl-short)

Challenge:	a recent study finds that commonsense reasoning is not always universal and can leave disabled people behind . a case study of disabled people with long COVID shows that common sense is not universal .
Approach:	They investigate how datasets and models deal with disability in commonsense reasoning . they use annotations from disabled and non-disabled persons for ableism .
Outcome:	The proposed datasets have low sensitivity to human-detected ableism but still detect 5 to 25% of entries as ableist.

Copied to clipboard

Phichet Phuangrot, Natdanai Trintawat, Kanawat Vilasri, Yanapat Patcharawiwatpong, Pachara Boonsarngsuk, Nat Pavasant, Ekapol Chuangsuwanich

Machine Translation

Challenge:	MEET-MR provides a comprehensive benchmark for evaluating English–Thai machine translation systems.
Approach:	They propose a benchmark for evaluating English–Thai machine translation systems . they use the Multidimensional Quality Metrics framework to provide fine-grained human judgements of translation quality.
Outcome:	The dataset covers nine domains providing linguistic and contextual diversity.

Evaluation of Deontic Conditional Reasoning in Large Language Models: The Case of Wason’s Selection Task (2026.eacl-short)

Copied to clipboard

Hirohiko Abe, Kentaro Ozeki, Risako Ando, Takanobu Morishita, Koji Mineshima, Mitsuhiro Okada

Confidence Leaps in LLM Reasoning: Early Stopping and Cross-Model Transfer (2026.eacl-short)

Challenge:	In humans, reasoning often performs well in domain specific settings, especially in normative rather than purely formal contexts.
Approach:	They propose a dataset that explicitly encodes deontic modality to systematically distinguish deontics from descriptive conditionals and analyze LLMs’ conditional reasoning under deontical rules.
Outcome:	The proposed dataset systematically distinguishes deontic from descriptive conditionals and examines LLMs’ conditional reasoning under deontics.

Copied to clipboard

Pavel Tikhonov, Ivan Oseledets, Elena Tutubalina

Exploring Fine-Tuning for In-Context Retrieval and Efficient KV-Caching in Long-Context Language Models (2026.eacl-short)

Challenge:	Large Language Models build confidence gradually during reasoning, but internal dynamics of how confidence evolves during this reasoning process remain poorly understood.
Approach:	They propose a model-agnostic early-stopping heuristic that halts generation upon detecting a "confidence leap" they argue that conviction is often reached in a discrete "moment of insight" they propose to train models without sacrificing accuracy .
Outcome:	The proposed model-agnostic heuristic reduces generation time without sacrificing accuracy and significantly reduces the generation time.

Copied to clipboard

Francesco Maria Molfese, Momchil Hardalov, Rexhina Blloshmi, Bill Byrne, Adrià de Gispert

Post-ASR Correction in Hindi: Comparing Language Models and Large Language Models in Low-Resource Scenarios (2026.eacl-short)

Challenge:	Long-Context Language Models (LCLMs) can encode entire document collections, offering a strong alternative to retrieval-augmented generation (RAG).
Approach:	They propose to use LCLMs to encode documents with context windows of millions of tokens to improve their performance.
Outcome:	The proposed training strategies improve long-context performance and their robustness under compression techniques.

Copied to clipboard

Rishabh Kumar, Amrith Krishna, Ganesh Ramakrishnan, Preethi Jyothi

CHiRPE: A Step Towards Real-World Clinical NLP with Clinician-Oriented Model Explanations (2026.eacl-short)

Challenge:	Automatic Speech Recognition (ASR) systems for low-resource languages produce erroneous transcripts due to limited annotated data and linguistic complexity.
Approach:	They compare language models and large language models for post-ASR correction in Hindi . they observe a scaling trend under zero-shot ICL where mid-sized LLMs degrade performance before marginal recovery at extreme scales.
Outcome:	The proposed model outperforms larger models in both fine-tuning and in-context learning settings.

Copied to clipboard

Stephanie Fong, Zimu Wang, Guilherme C Oliveira, Xiangyu Zhao, Yiwen Jiang, Jiahe Liu, Beau-Luke Colton, Scott W. Woods, Martha Shenton, Barnaby Nelson, Zongyuan Ge, Dominic Dwyer

Explainability & Interpretability in NLP

Challenge:	Psychotic disorders are a major contributor to the global health burden due to their relatively high mortality risk.
Approach:	They propose an NLP pipeline that takes semi-structured clinical interviews to predict psychosis risk and generate novel SHAP explanation formats.
Outcome:	The proposed pipeline outperforms baseline models and achieves 90% accuracy across three BERT variants.

LLMs Know More About Numbers than They Can Say (2026.eacl-short)

Copied to clipboard

Fengting Yuchi, Li Du, Jason Eisner

Numerical Reasoning

Challenge:	Large language models (LLMs) are increasingly used in mathematical, scientific, financial and engineering domains.
Approach:	They probe the hidden states of several smaller open-source LLMs to find out how big they are .
Outcome:	The proposed model improves verbalized accuracy by 3.22% over base models.

On the Mathematical Relationship Between Layer Normalization and Dynamic Activation Functions (2026.eacl-short)

Copied to clipboard

Felix Stollenwerk

Semantic Token Clustering for Efficient Uncertainty Quantification in Large Language Models (2026.eacl-short)

Challenge:	Layer normalization (LN) is an essential component of modern neural networks.
Approach:	They propose a dynamic activation function called Dynamic Tanh which is based on the LN variant RMSNorm and decouples in derivative space.
Outcome:	The proposed function reproduces the normalization effect on outliers more accurately than RMSNorm by using a well-defined decoupling procedure in derivative space and an approximation.

Copied to clipboard

Qi Cao, Andrew Gambardella, Takeshi Kojima, Yutaka Matsuo, Yusuke Iwasawa

Becoming Experienced Judges: Selective Test-Time Learning for Evaluators (2026.eacl-short)

Challenge:	Large language models have limited truthfulness and tendency toward overconfidence constrain reliability in factual tasks.
Approach:	They propose an efficient method that leverages semantic information encoded in LLMs to quantify uncertainty.
Outcome:	The proposed method achieves comparable performance to baselines while significantly reducing computational overhead.

Copied to clipboard

Seungyeon Jwa, Daechul Ahn, Reokyoung Kim, Dongyeop Kang, Jonghyun Choi

Statistical Foundations of DIME: Risk Estimation for Practical Index Selection (2026.eacl-short)

Challenge:	Large language models and visionlanguage models are increasingly used as automatic evaluators.
Approach:	They propose a framework that allows evaluators to improve sequentially at inference time without additional training or external signals.
Outcome:	The proposed framework outperforms strong baselines in two pairwise comparisons.

Copied to clipboard

Giulio D'Erasmo, Cesare Campagnano, Antonio Mallia, Pierpaolo Brutti, Nicola Tonellotto, Fabrizio Silvestri