Austria

Grounding the reasoning ability of large language models (LLMs) for embodied tasks is challenging due to the complexity of the physical world. Especially, LLM planning for multi-agent collaboration requires communication of agents or credit assignment as the feedback to re-adjust the proposed plans and achieve effective coordination. However, existing methods that overly rely on physical verification or self-reflection suffer from excessive and inefficient querying of LLMs. In this paper, we propose a novel framework for multi-agent collaboration that introduces Reinforced Advantage feedback (ReAd) for efficient self-refinement of plans. Specifically, we perform critic regression to learn a sequential advantage function from LLM-planned data, and then treat the LLM planner as an optimizer to generate actions that maximize the advantage function. It endows the LLM with the foresight to discern whether the action contributes to accomplishing the final task. We provide theoretical analysis by extending advantage-weighted regression in reinforcement learning to multi-agent systems. Experiments on Overcooked-AI and a difficult variant of RoCoBench show that ReAd surpasses baselines in success rate, and also significantly decreases the interaction steps of agents and query rounds of LLMs, demonstrating its high efficiency for grounding LLMs. More results are given at https://read-llm.github.io/.

ACL 2025

Towards Efficient LLM Grounding for Embodied Multi-Agent Collaboration

llm planning

multi-agent collaboration

embodied ai

large language models

### 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)

poster

Identifying linguistic bias in text requires understanding what is said and what is meant. This requires going beyond what is being asserted directly, and determining what is presupposed. Large language models (LLMs) represent a potential automatic approach for identifying presupposed content, but it is unknown how well LLM judgments correspond to human judgments. Further, LLMs may exhibit their own biases in determining what is presupposed. To study this empirically, we prompt multiple LLMs to make presupposition judgments for texts of varying domains from three different human-labeled datasets. We calculate the agreement between LLMs and human raters, and find several linguistic factors associated with fluctuations in human-model agreement. We also observe discrepancies in agreement scores that indicate potential biases from LLMs, related to gender and political ideology.

Measuring Bias and Agreement in Large Language Model Presupposition Judgments

Effective nurse communication is vital for high-quality patient care, yet standardized patient (SP) simulations are costly and inflexible, and virtual patient (VP) systems often fail to adapt dynamically to trainees' communication skills. When a trainee responds ineffectively, the VP should escalate hostility or become uncooperative, yet most systems lack this adaptive capability. To address this, we introduce Adaptive-VP, an LLM-based framework that flexibly adjusts VP dialogue. Our framework includes a pipeline for constructing clinically grounded yet adaptable VP cases and implements a modular structure for evaluating trainee communication and dynamically adjusting VP responses, while ensuring learner safety. Specifically, we applied and validated Adaptive-VP by simulating challenging patient conversations. Automated evaluation of the corpus from practicing nurses showed that our evaluation mechanism mirrored real-world communication efficiency levels. Human evaluation with expert nurses further confirmed that Adaptive-VP generates more realistic and natural interactions than existing approaches, demonstrating its potential as a scalable and effective tool for nursing communication.

Adaptive-VP: A Framework for LLM-Based Virtual Patients that Adapts to Trainees' Dialogue to Facilitate Nurse Communication Training

Logical reasoning is fundamental to intelligent systems. Large language models (LLMs) have demonstrated promise in natural language (NL) reasoning, especially with techniques like chain-of-thought (CoT) prompting. Neuro-symbolic methods like Logic-LM and LINC further enhance performance on challenging datasets FOLIO and AR-LSAT by integrating formalization with LLMs and symbolic solvers, and possibly refinement with LLMs. However, these methods still struggle with the accurate formalization of complex NL problems. In this paper, we introduce LTRAG, a framework to enhance autoformalization and self-refinement for logical reasoning with Retrieval-Augmented Generation (RAG), by building knowledge bases of thought-guided examples. Experimental results on FOLIO and AR-LSAT show that LTRAG consistently outperforms Logic-LM and LINC across different models. On GPT-4 and AR-LSAT, it achieves an accuracy gain of 13% over Logic-LM.

LTRAG: Enhancing autoformalization and self-refinement for logical reasoning with Thought-Guided RAG

Retrieval-augmented language models (RALMs) aim to incorporate external knowledge to address the issues of factual hallucination and knowledge obsolescence faced by large language models (LLMs). Inevitably, the retrieved passages based on similarity search may be irrelevant to the given question, and the aggregation of these passages can confuse the model to give a correct answer. To improve the performance of RALM in such conditions, we propose layer-knowledge guided attention for RALMs, which harnesses the layer-wise knowledge of LLMs to optimize per-layer attention on useful passages, making the model pay attention to the most relevant content and ignore irrelevant ones. Specifically, we first systematically study LLM's attention patterns and their relationship with the accuracy of RALM responses, where middle-focus attentions play a crucial role in selectively gathering relevant information. Based on this, a layer-wise passage estimator leverages the varied knowledge encoded across LLM layers to assess not only passage relevance scores but also associated confidences. Finally, a relevance-aware passage fusion enables selective attention to relevant passages, mitigating distractibility and positional bias of causal attention. Experiments show that our method outperforms existing methods on RALM benchmarks.

Making RALM Robust to Irrelevant Contexts via Layer Knowledge Guided Attention

Deaf or Hard-of-Hearing (DHH) individuals use Large Language Models (LLMs) in unique ways and request to incorporate sign language grammar and Deaf culture during the training of these models, in addition to video-based sign language input capabilities. Yet, developers of instruct-tuned LLMs have not paid attention to these requests. Instead, special translation models are developed for sign languages (SLs), diminishing the needs of signers to a simple lack of communication between hearing and Deaf communities. In this paper, we take an orthogonal approach to these traditional methods of studying SLs. To meet the requests of Deaf users of LLMs, we look at the sign language processing (SLP) algorithm from a theoretical lens, then introduce the first text-based and multimodal LLMs. We propose new prompting and fine-tuning strategies for text-based and multimodal SLP, incorporating sign linguistic rules and conventions. We test the generalization of these models to other SLP tasks, showing LLMs can process signs while still being adept at spoken language tasks. Our code and model checkpoints will be open-source. We will update our model suite as newer open-source LLMs, datasets, and SLP tasks become available.

Large Language Models as Sign Language Interfaces: Mitigating the Requests of Deaf Users of LLMs in a Hearing-Centric World

Chain-of-thought (CoT) prompting has been
widely adopted to enhance the reasoning ca-
pabilities of large language models (LLMs).
However, the effectiveness of CoT reasoning is
inconsistent across tasks with different reason-
ing types. This work presents a novel perspec-
tive to understand CoT behavior through the
lens of confirmation bias in cognitive psychol-
ogy. Specifically, we examine how model in-
ternal beliefs, approximated by direct question-
answering probabilities, affect both reasoning
generation (Q -> R) and reasoning-guided an-
swer prediction (QR -> A) in CoT. By decom-
posing CoT into a two-stage process, we con-
duct a thorough correlation analysis in model
beliefs, rationale attributes, and stage-wise per-
formance. Our results provide strong evidence
of confirmation bias in LLMs, such that model
beliefs not only skew the reasoning process but
also influence how rationales are utilized for
answer prediction. Furthermore, the interplay
between task vulnerability to confirmation bias
and the strength of beliefs also provides expla-
nations for CoT effectiveness across reasoning
tasks and models. Overall, this study provides a
valuable insight for the needs of better prompt-
ing strategies that mitigate confirmation bias to
enhance reasoning performance. Code is avail-
able at https://github.com/yuewan2/biasedcot.

Unveiling Confirmation Bias in Chain-of-Thought Reasoning

Representational harms are widely recognized among fairness-related harms caused by generative language systems. However, their definitions are commonly under-specified. We present a framework, grounded in speech act theory (Austin, 1962), that conceptualizes representational harms caused by generative language systems as the perlocutionary effects (i.e., real-world impacts) of particular types of illocutionary acts (i.e., system behaviors). Building on this argument and drawing on relevant literature from linguistic anthropology and sociolinguistics, we provide new definitions for stereotyping, demeaning, and erasure. We then use our framework to develop a granular taxonomy of illocutionary acts that cause representational harms, going beyond the high-level taxonomies presented in previous work. We also discuss the ways that our framework and taxonomy can support the development of valid measurement instruments. Finally, we demonstrate the utility of our framework and taxonomy via a case study that engages with recent conceptual debates about what constitutes a representational harm and how such harms should be measured.

Taxonomizing Representational Harms using Speech Act Theory

We propose a novel scaling law for general-purpose decoder-only language models (LMs) trained on multilingual data, tackling the problem of balancing languages during multilingual pretraining. A primary challenge in studying multilingual scaling is the difficulty of analyzing individual language performance due to cross-lingual transfer. To tackle this, we shift the focus from individual languages to language families. We introduce and validate a hypothesis that the test cross-entropy loss for each language family is determined solely by its own sampling ratio, independent of other languages in the mixture. This insight simplifies the complexity of multilingual scaling and make the analysis scalable to an arbitrary number of languages. Building on this hypothesis, we derive a power-law relationship that links performance with dataset size, model size and sampling ratios. This relationship enables us to predict performance across various combinations of the above three quantities, and derive the optimal sampling ratios at different model scales. To demonstrate the effectiveness and accuracy of our proposed scaling law, we perform a large-scale empirical study, training more than 100 models on 23 languages spanning 5 language families. Our experiments show that the optimal sampling ratios derived from small models (85M parameters) generalize effectively to models that are several orders of magnitude larger (1.2B parameters), offering a resource-efficient approach for multilingual LM training at scale.

Scaling Laws for Multilingual Language Models

Long-context understanding is crucial for many NLP applications, yet transformers struggle with efficiency due to the quadratic complexity of self-attention. Sparse attention methods alleviate this cost but often impose static, predefined masks, failing to capture heterogeneous attention patterns. This results in suboptimal token interactions, limiting adaptability and retrieval accuracy in long-sequence tasks. This work introduces a dynamic sparse attention mechanism that assigns adaptive masks at the attention-map level, preserving heterogeneous patterns across layers and heads. Unlike existing approaches, our method eliminates the need for fine-tuning and predefined mask structures while maintaining computational efficiency. By learning context-aware attention structures, it achieves high alignment with full-attention models, ensuring minimal performance degradation while reducing memory and compute overhead. This approach provides a scalable alternative to full attention, enabling the practical deployment of large-scale Large Langue Models (LLMs) without sacrificing retrieval performance.

DAM: Dynamic Attention Mask for Long-Context Large Language Model Inference Acceleration

With the advent of LLMs and variants, a flurry of research has emerged, analyzing the performance of such models across an array of tasks. While most studies focus on evaluating the capabilities of state-of-the-art (SoTA) Vision Language Models (VLMs) through task accuracy (e.g., visual question answering, grounding), our work explores the related but complementary aspect of consistency – the ability of a VLM to produce semantically similar or identical responses to semantically similar queries. We note that consistency is a fundamental prerequisite (necessary but not sufficient condition) for robustness and trust in VLMs. Armed with this perspective, we propose the MM-Rbenchmark, which allows us to analyze performance, in terms of consistency and accuracy, of SoTA VLMs on three tasks: Question Rephrasing, Image Restyling, and Context Reasoning. Our analysis reveals that consistency does not always align with accuracy, indicating that models with higher accuracy are not necessarily more consistent, and vice versa. Furthermore, we propose a simple yet effective mitigation strategy in the form of an adapter module trained to minimize inconsistency across prompts. With our proposed strategy, we are able to achieve absolute improvements of 5.7% and 12.5%, on average on widely used VLMs such as BLIP-2 and LLaVa 1.5M in terms of consistency over their existing counterparts.

Downloads

Next from ACL 2025

Measuring Bias and Agreement in Large Language Model Presupposition Judgments

Similar lecture

Collaborating Action by Action: A Multi-agent LLM Framework for Embodied Reasoning

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES