United States

Large language models (LLMs) have achieved remarkable success across a wide range of natural language processing tasks, yet their performance remains heavily biased toward high-resource languages. Tibetan, despite its cultural significance and large speaker population, is still substantially underrepresented. In this work, we present a comprehensive pipeline for advancing Tibetan language modeling through large-scale data curation and continual pre-training. We construct a 72 GB high-quality Tibetan corpus, the largest to date, and adapt Qwen2.5-7B through balanced multilingual continual pre-training with Tibetan, Chinese, and English, followed by multilingual instruction tuning. To further scale capacity efficiently, we extend the dense model to a 50B-A10B Mixture-of-Experts architecture. Due to the absence of standardized Tibetan benchmarks, we build multiple evaluation datasets via high-quality translation and human verification. Experimental results show that both dense and MoE models consistently outperform existing open-source and Tibetan-focused models of similar scale across diverse tasks. Our work advances Tibetan-centric LLM research and provides transferable insights for extending LLMs to other low-resource languages. We will release the model weights, evaluation benchmarks, and detailed data processing documentation in the follow-up.

ACL 2026

From Curated Data to Scalable Models: Continual Pre-training of Dense and MoE Large Language Models for Tibetan

It is my great pleasure and honor to welcome you to the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), held in San Diego, California, USA, July 2–7, 2026. ACL 2026 continues our fields 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.

It is exciting to see the many developments in the field reflected at our conference, such as the advancement of large language models (born out of our community) into many application directions. This is not without challenges, partly due to unsolved problems, and partly due to misuse of the technology. All this is reflected well at ACL 2026. The conference also reflects that the field is not only about advancing technology, but also the science of fundamentally understanding language and how it is use by many different speaker communities. <br>
[Full Message from General Chair Philipp Koehn](https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing) <br><br><br>
<html><button style="display: inline-flex; align-items: center; justify-content: center; white-space: nowrap; border-radius: 9999px; font-weight: bold; background: #7c3aed; color: white; font-family: 'Space Grotesk', sans-serif; height: 40px; font-size: 16px; padding: 0 20px; border: none; cursor: pointer" onclick="window.open('https://drive.google.com/file/d/1QiJFOmMjvDgsDo3i1XzlQnImxe3RU_Cy/view?usp=sharing')">Conference Handbook</button></html>
<br><br><br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to ACL 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://2026.aclweb.org/registration/) first.

**Online Registration Form**: https://acl.swoogo.com/acl2026

Registration Required

The 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026) will take place in San Diego, California, from July 2 - 7, 2026.

technical paper

Code retrieval is vital to modern software engineering as it boosts reuse and speeds up debugging. However, current benchmarks primarily emphasize functional relevance while neglecting code quality. To address this gap, we introduce CoQuIR, the first large-scale, multilingual benchmark specifically designed to evaluate quality-aware code retrieval across four critical dimensions: correctness, efficiency, security, and maintainability. CoQuIR includes fine-grained quality annotations over 42,725 queries and 134,907 code snippets in 11 programming languages. Evaluating 23 retrievers (both open-source and proprietary) shows that even state-of-the-art models often fail to separate buggy or insecure code from robust counterparts. We further investigate methods for explicitly training retrievers to recognize code quality, demonstrating that quality-aware metrics can be improved without loss of semantic relevance; downstream code generation benefits from these gains. CoQuIR underscores the importance of embedding quality signals into retrieval systems as a crucial component for more trustworthy developer tools.

CoQuIR: A Comprehensive Benchmark for Code Quality-Aware Information Retrieval

Rubric-augmented verification guides reward models with explicit evaluation criteria, yielding more reliable judgments than single-model verification.
However, most existing methods require costly rubric annotations, limiting scalability.
Moreover, we find that rubric generation is vulnerable to a failure of cooperation; low-quality rubrics actively mislead reward models rather than help. 
Inspired by the principle of cooperative communication, we propose Cooperative yet Critical reward modeling (C2), a framework that significantly improves reward model judgments by having the reward model critically collaborate with a rubric generator trained solely from binary preferences.
In C2, we synthesize helpful and misleading rubric pairs by measuring how each rubric shifts the reward model toward or away from the correct preference. 
Using these contrastive pairs, we train a cooperative rubric generator to propose helpful rubrics, and a critical verifier to assess rubric validity before making its judgment, following only rubrics it deems helpful at inference time. 
C2 outperforms reasoning reward models trained on the same binary preferences, with gains of up to 6.5 points on RM-Bench and 6.0 points length-controlled win rate on AlpacaEval 2.0. 
Without external rubric annotations, C2 enables an 8B reward model to match performance achieved with rubrics from a 4$\times$ larger model.
Overall, our work demonstrates that eliciting deliberate cooperation in rubric-augmented verification makes reward models more trustworthy in a scalable way.

C2: Scalable Rubric-Augmented Reward Modeling from Binary Preferences

While bias in large language models (LLMs) is well-studied, similar concerns in vision-language models (VLMs) have received comparatively less attention. Existing VLM bias studies often focus on portrait-style images and gender-occupation associations, overlooking broader and more complex social stereotypes and their implied harm. This work introduces Vignette, a large-scale VQA benchmark with 30M+ images for evaluating bias in VLMs through a question-answering framework spanning four directions: factuality, perception, stereotyping, and decision making. Beyond narrowly-centered studies, we assess how VLMs interpret identities in contextualized settings, revealing how models make trait and capability assumptions and exhibit patterns of discrimination. Drawing from social psychology, we examine how VLMs connect visual identity cues to trait and role-based inferences, encoding social hierarchies, through biased selections. Our findings uncover subtle, multifaceted, and surprising stereotypical patterns, offering insights into how VLMs construct social meaning from inputs.

VIGNETTE: Socially Grounded Bias Evaluation for Vision-Language Models

Culture serves as a fundamental determinant of human affective processing and profoundly shapes how individuals perceive and interpret emotional stimuli. Despite this intrinsic link extant evaluations regarding cultural alignment within Large Language Models primarily prioritize declarative knowledge such as geographical facts or established societal customs. These benchmarks remain insufficient to capture the subjective interpretative variance inherent to diverse sociocultural lenses. To address this limitation, we introduce CEDAR, a multimodal benchmark constructed entirely from scenarios capturing **C**ulturally **E**licited **D**istinct **A**ffective **R**esponses. To construct CEDAR, we implement a novel pipeline that leverages LLM-generated provisional labels to isolate instances yielding cross-cultural emotional distinctions, and subsequently derives reliable ground-truth annotations through rigorous human evaluation. The resulting benchmark comprises 10,962 instances across seven languages and 14 fine-grained emotion categories, with each language including 400 multimodal and 1,166 text-only samples. Comprehensive evaluations of 17 representative multilingual models reveal a dissociation between language consistency and cultural alignment, demonstrating that culturally grounded affective understanding remains a significant challenge for current models.

Tears or Cheers? Benchmarking LLMs via Culturally Elicited Distinct Affective Responses

Large language models (LLMs) operate in two fundamental learning modes – fine-tuning (FT) and in-context learning (ICL) – raising key questions about which mode yields greater language proficiency and whether they differ in their inductive biases. Prior studies comparing FT and ICL have yielded mixed and inconclusive results due to inconsistent experimental setups. To enable a rigorous comparison, we propose a formal language learning task – offering precise language boundaries, controlled string sampling, and no data contamination – and introduce a discriminative test for language proficiency, where an LLM succeeds if it assigns higher generation probability to in-language strings than to out-of-language strings.

Empirically, we find that: (a) FT has greater language proficiency than ICL on in-distribution generalization, but both perform equally well on out-of-distribution generalization. (b) Their inductive biases, measured by the correlation in string generation probabilities, are similar when both modes partially learn the language but diverge at higher proficiency levels. (c) Unlike FT, ICL performance differs substantially across models of varying sizes and families and is sensitive to the token vocabulary of the language. Thus, our work demonstrates the promise of formal languages as a controlled testbed for evaluating LLMs, behaviors that are difficult to isolate in natural language datasets. Our source code is available at https://github.com/bishwamittra/formallm.

Fine-tuning vs. In-context Learning in Large Language Models: A Formal Language Learning Perspective

The growing scale of online misinformation urgently demands Automated Fact-Checking (AFC). Existing benchmarks for evaluating AFC systems, however, are largely limited in terms of task scope, modalities, domain, language diversity, realism, or coverage of misinformation types. Critically, they are static, thus subject to data leakage as their claims enter the pretraining corpora of LLMs. As a result, benchmark performance no longer reliably reflects the actual ability to verify claims.
We introduce Verified Theses and Statements (VeriTaS), the first dynamic benchmark for multimodal AFC, designed to remain robust under ongoing large-scale pretraining of foundation models. VeriTaS currently comprises 25,000 real-world claims from 104 professional fact-checking organizations across 54 languages, covering textual and audiovisual content. Claims are added quarterly via a fully automated seven-stage pipeline that normalizes claim formulation, retrieves original media, and maps heterogeneous expert verdicts to a novel, standardized, and disentangled scoring scheme with textual justifications.
Through human evaluation, we demonstrate that the automated annotations closely match human judgments.
We commit to updating VeriTaS in the future, establishing a leakage-resistant benchmark, supporting meaningful AFC evaluation in the era of rapidly evolving foundation models.
The code and data are publicly available under https://veritas.mai.informatik.tu-darmstadt.de.

VeriTaS: The First Dynamic Benchmark for Multimodal Automated Fact-Checking

Distantly Supervised Relation Extraction (DSRE) remains a long-standing challenge in NLP, where models must learn from noisy bag-level annotations while making sentence-level predictions. While existing state-of-the-art (SoTA) DSRE models rely on task-specific training, their integration with in-context learning (ICL) using large language models (LLMs) remains underexplored. A key challenge is that the LLM may not learn relation semantics correctly, due to noisy annotation.

In response, we propose $HYDRE$ -- $HY$brid $D$istantly Supervised $R$elation $E$xtraction framework. It first uses a trained DSRE model to identify the top-$k$ candidate relations for a given test sentence, then uses a novel dynamic exemplar retrieval strategy that extracts reliable, sentence-level exemplars from training data, which are then provided in LLM prompt for outputting the final relation(s).

We further extend $HYDRE$ to cross-lingual settings for RE in low-resource languages. Using available English DSRE training data, we evaluate all methods on English as well as a newly curated benchmark covering four diverse low-resource Indic languages - Oriya, Santali, Manipuri, and Tulu. $HYDRE$ achieves up to 20 F1 point gains in English and, on average, 17 F1 points on Indic languages over prior SoTA DSRE models and naive prompting baselines. Detailed ablations exhibit $HYDRE$'s efficacy compared to other prompting strategies.

Combining Distantly Supervised Models with In Context Learning for Monolingual and Cross-Lingual Relation Extraction

Task-oriented proactive dialogue agents play a pivotal role in recruitment, particularly for steering conversations towards specific business outcomes, such as acquiring social-media contacts for private-channel conversion. Although supervised fine-tuning and reinforcement learning have proven effective for training such agents, their performance is heavily constrained by the scarcity of high-quality, goal-oriented domain-specific training data. To address this challenge, we propose $\textbf{SimRPD}$, a three-stage framework for training recruitment proactive dialogue agents. First, we develop a high-fidelity user simulator to synthesize large-scale conversational data through multi-turn online dialogue. Then we introduce a multi-dimensional evaluation framework based on $\textbf{Chain-of-Intention (CoI)}$ to comprehensively assess the simulator and effectively select high-quality data, incorporating both global-level and instance-level metrics. Finally, we train the recruitment proactive dialogue agent on the selected dataset. Experiments in a real-world recruitment scenario demonstrate that SimRPD outperforms existing simulator-based data selection strategies, highlighting its practical value for industrial deployment and its potential applicability to other business-oriented dialogue scenarios.

SimRPD: Optimizing Recruitment Proactive Dialogue Agents through Simulator-Based Data Evaluation and Selection

Reasoning LLMs often spend substantial tokens on long intermediate reasoning traces (e.g., chain-of-thought) when solving new problems. We propose to summarize and store reusable reasoning skills distilled from extensive deliberation and trial-and-error exploration, and to retrieve these skills at inference time to guide future reasoning. Unlike the prevailing \emph{reasoning from scratch} paradigm, our approach first recalls relevant skills for each query, helping the model avoid redundant detours and focus on effective solution paths. We evaluate our method on coding and mathematical reasoning tasks, and find that it significantly reduces reasoning tokens while improving overall performance. The resulting lower per-request cost indicates strong practical and economic potential for real-world deployment.

Thinking with Reasoning Skills: Fewer Tokens, More Accuracy

Chain-of-Thought (CoT) in large language models (LLMs) has been widely debated in terms of whether it faithfully reflects an internal reasoning process of models. Parametric faithfulness is a recently proposed metric that uses unlearning to assess whether a model encodes parametric beliefs corresponding to a reasoning chain. This paper refines this metric by accounting for the unintended artifacts of unlearning. We introduce control tasks that unlearn irrelevant knowledge and word-shuffled content and show that these control tasks can measure parametric faithfulness, suggesting the non-negligible effect of unlearning. We also found that control tasks help explain the significant variations in parametric faithfulness observed across different model sizes and CoT lengths. We conclude that the effects of unlearning should be considered when measuring parametric faithfulness.

Downloads

Next from ACL 2026

CoQuIR: A Comprehensive Benchmark for Code Quality-Aware Information Retrieval

Similar lecture

Scaling Sentence Similarity for Classical Tibetan with Automatic Annotations

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from ACL 2026

CoQuIR: A Comprehensive Benchmark for Code Quality-Aware Information Retrieval

Similar lecture

Scaling Sentence Similarity for Classical Tibetan with Automatic Annotations

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads