Austria

Large Language Models (LLMs) are valued for their strong performance across various tasks, but they also produce inaccurate or misleading outputs. Uncertainty Estimation (UE) quantifies the model&#39;s confidence and helps users assess response reliability. However, existing UE methods have not been thoroughly examined in scenarios like Retrieval-Augmented Generation (RAG), where the input prompt includes non-parametric knowledge. This paper shows that current UE methods cannot reliably assess correctness in the RAG setting. We further propose an axiomatic framework to identify deficiencies in existing methods and guide the development of improved approaches. Our framework introduces five constraints that an effective UE method should meet after incorporating retrieved documents into the LLM&#39;s prompt. Experimental results reveal that no existing UE method fully satisfies all the axioms, explaining their suboptimal performance in RAG. We further introduce a simple yet effective calibration function based on our framework, which not only satisfies more axioms than baseline methods but also improves the correlation between uncertainty estimates and correctness.

ACL 2025

Axiomatic Analysis of Uncertainty Estimation For Retrieval Augmented Generation

axiomatic thinking

retrieval augmented generation

uncertainty estimation

### 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)

Large Language Models (LLMs) are valued for their strong performance across various tasks, but they also produce inaccurate or misleading outputs. Uncertainty Estimation (UE) quantifies the model's confidence and helps users assess response reliability. However, existing UE methods have not been thoroughly examined in scenarios like Retrieval-Augmented Generation (RAG), where the input prompt includes non-parametric knowledge. This paper shows that current UE methods cannot reliably assess correctness in the RAG setting. We further propose an axiomatic framework to identify deficiencies in existing methods and guide the development of improved approaches. Our framework introduces five constraints that an effective UE method should meet after incorporating retrieved documents into the LLM's prompt. Experimental results reveal that no existing UE method fully satisfies all the axioms, explaining their suboptimal performance in RAG. We further introduce a simple yet effective calibration function based on our framework, which not only satisfies more axioms than baseline methods but also improves the correlation between uncertainty estimates and correctness.

poster

Misinformation is a global issue that shapes public discourse, influencing opinions and decision-making across various domains. While automated fact-checking (AFC) has become essential in combating misinformation, most work in multilingual settings has focused on claim verification rather than generating explanatory verdicts, leaving a gap in AFC resources beyond English. To address this, we introduce EuroVerdict, a multilingual dataset designed for verdict generation, covering eight European languages. Developed in collaboration with professional fact-checkers, the dataset includes claims, manually written verdicts, and supporting evidence, consisting of fact-checking articles and additional secondary sources. We evaluate EuroVerdict with Llama-3.1-8B-Instruct on verdict generation under different settings, varying the prompt language, input article language, and training approach. Our results show that fine-tuning consistently improves performance, with models fine-tuned on original-language articles achieving the highest scores in both automatic and human evaluations. Using articles in a different language than the claim slightly lowers performance; however, pairing them with language-specific prompts improves results. Zero-shot and Chain of Thought configurations perform worse, reinforcing the benefits of fine-tuning for multilingual verdict generation.

EuroVerdict: A Multilingual Dataset for Verdict Generation Against Misinformation

Large language models (LLMs) encode vast amounts of world knowledge acquired via training on large web-scale datasets crawled from the internet. However, the datasets used to train the LLMs typically exhibit a geographical bias towards English-speaking Western countries. This results in LLMs producing biased or hallucinated responses to queries that require answers localized to other geographical regions. In this work, we introduce a new benchmark named LoFTI (Localization and Factuality Transfer to Indian Locales) that can be used to evaluate an LLM's contextual localization and factual text transfer capabilities. LoFTI consists of factual statements about entities in source and target locations; the source locations are spread across the globe and the target locations are all within India with varying degrees of hyperlocality (country, states, cities). The entities span a wide variety of categories. We use LoFTI to evaluate Mixtral, Llama3.3-70B, GPT-4 and two other Mixtral-based approaches well-suited to the task of localized factual transfer. We demonstrate that LoFTI is a high-quality evaluation benchmark and all the models, including GPT-4, produce skewed results across varying levels of hyperlocality.

LoFTI: Localization and Factuality Transfer to Indian Locales

Large Language Models (LLMs) have shown impressive potential in clinical question answering (QA), with Retrieval Augmented Generation (RAG) emerging as a leading approach for ensuring the factual accuracy of model responses. However, current automated RAG metrics perform poorly in clinical and conversational use cases. Using clinical human evaluations of responses is expensive, unscalable, and not conducive to the continuous iterative development of RAG systems. To address these challenges, we introduce ASTRID - an Automated and Scalable TRIaD for evaluating clinical QA systems leveraging RAG - consisting of three metrics: Context Relevance (CR), Refusal Accuracy (RA), and Conversational Faithfulness (CF). Our novel evaluation metric, CF, is designed to better capture the faithfulness of a model’s response to the knowledge base without penalising conversational elements. Additionally, our metric RA captures the refusal to address questions outside of the system’s scope of practice. To validate our triad, we curate a dataset of over 200 real-world patient questions posed to an LLM-based QA agent during surgical follow-up for cataract surgery - the highest volume operation in the world - augmented with clinician-selected questions for emergency, and clinical and non-clinical out-of-domain scenarios. We demonstrate that CF predicts human ratings of faithfulness more accurately than existing definitions in conversational settings. Furthermore, using eight different LLMs, we demonstrate that the three metrics can closely agree with human evaluations, highlighting the potential of these metrics for use in LLM-driven automated evaluation pipelines. Finally, we show that evaluation using our triad of CF, RA, and CR exhibits alignment with clinician assessment for inappropriate, harmful, or unhelpful responses. We also publish the prompts and datasets for these experiments, providing valuable resources for further research and development.

ASTRID - An Automated and Scalable TRIaD for the Evaluation of RAG-based Clinical Question Answering Systems

This study investigates the condition under which Transformer-based language models (LMs) identify and distinguish entities, such as individuals and organizations, with their internal representations, focusing on the many-to-many correspondence between entities and their mentions. We first formulate two problems, ambiguity and variability of entity mentions, and propose a framework analogous to clustering quality metrics to measure how well the models resolve these issues. Our experiments examine four Transformer-based autoregressive models, showing that they effectively identify and distinguish entities with metrics analogous to precision and recall exceeding 0.8. Further analysis reveals that entity-related information is compactly represented in a low-dimensional linear subspace at early LM layers and provides potential interpretations of these findings through the lens of isomorphism between LM representations and entity-centric knowledge structures in the real world.

On Entity Identification in Language Models

Handling ambiguity and underspecification is an important challenge in natural language interfaces, particularly for tasks like text-to-SQL semantic parsing. We propose a modular approach that resolves ambiguity using natural language interpretations before mapping these to logical forms (e.g., SQL queries). Although LLMs excel at parsing unambiguous utterances, they show strong biases for ambiguous ones, typically predicting only preferred interpretations. We constructively exploit this bias to generate an initial set of preferred disambiguations and then apply a specialized infilling model to identify and generate missing interpretations. To train the infilling model, we introduce an annotation method that uses SQL execution to validate different meanings. Our approach improves interpretation coverage and generalizes across datasets with different annotation styles, database structures, and ambiguity types.

Disambiguate First Parse Later: Generating Interpretations for Ambiguity Resolution in Semantic Parsing

Automatic medical coding has the potential to ease documentation and billing processes. For this task, transparency plays an important role for medical coders and regulatory bodies, which can be achieved using explainability methods. However, the evaluation of these approaches has been mostly limited to short text and binary settings due to a scarcity of annotated data. Recent efforts by Cheng et al. (2023) have introduced the MDACE dataset, which provides a valuable resource containing code evidence in clinical records. In this work, we conduct an in-depth analysis of the MDACE dataset and perform plausibility evaluation of current explainable ICD coding systems from an applied perspective. With this, we contribute to a deeper understanding of automatic ICD coding and evidence extraction. Our findings reveal that ground truth evidence aligns with code descriptions to a certain degree. An investigation into state-of-the-art approaches shows a high overlap with ground truth evidence. We propose introduce match measures and highlight success and failure cases. Based on our findings, we provide recommendations for developing and evaluating explainable ICD coding systems.

The anatomy of evidence: An investigation into explainable ICD coding

Human readers can efficiently comprehend scrambled words, a phenomenon known as Typoglycemia, primarily by relying on word form; if word form alone is insufficient, they further utilize contextual cues for interpretation. While advanced large language models (LLMs) exhibit similar abilities, the underlying mechanisms remain unclear. To investigate this, we conduct controlled experiments to analyze the roles of word form and contextual information in semantic reconstruction and examine LLM attention patterns. Specifically, we first propose SemRecScore, a reliable metric to quantify the degree of semantic reconstruction, and validate its effectiveness. Using this metric, we study how word form and contextual information influence LLMs’ semantic reconstruction ability, identifying word form as the core factor in this process. Furthermore, we analyze how LLMs utilize word form and find that they rely on specialized attention heads to extract and process word form information, with this mechanism remaining stable across varying levels of word scrambling. This distinction between LLMs’ fixed attention patterns primarily focused on word form and human readers’ adaptive strategy in balancing word form and contextual information provides insights into enhancing LLM performance by incorporating human-like, context-aware mechanisms. Code is available on: https://github.com/Aurora-cx/TypoLLM.

Word Form Matters: LLMs’ Semantic Reconstruction under Typoglycemia

Model knowledge editing enables the efficient correction of erroneous information and the continuous updating of outdated knowledge within language models. While existing research has demonstrated strong performance in single-instance or few-instance sequential editing and one-time massive editing scenarios, the batched sequential editing paradigm remains a significant challenge. The primary issue lies in the model’s tendency to gradually forget previously edited knowledge and become increasingly unstable after multiple iterations of batched editing. 
To address these challenges, we propose **SeqMMR**, an enhanced framework for batched sequential knowledge editing that leverages **Seq**uential **M**odel **M**erging and a model **R**outer. Our approach iteratively merges parameters from current batch-edited models with those of their predecessors, ensuring that newly emerging knowledge is integrated while mitigating the forgetting of previously edited knowledge. Furthermore, the model router directs queries unrelated to the edited knowledge to an unedited model backup, preventing unintended alterations in model predictions. Extensive experiments across various datasets demonstrate that our approach effectively mitigates knowledge forgetting, improves performance across all previous batches, and better preserves the model's general capabilities.

SeqMMR: Sequential Model Merging and LLM Routing for Enhanced Batched Sequential Knowledge Editing

We present a novel pipeline, ReflectEvo, to demonstrate that small language models (SLMs) can enhance meta introspection through reflection learning. This process iteratively generates self-reflection for self-training, fostering a continuous and self-evolving process. Leveraging this pipeline, we construct ReflectEvo-460k, a large-scale, comprehensive, self-generated reflection dataset with broadened instructions and diverse multi-domain tasks. Building upon this dataset, we demonstrate the effectiveness of reflection learning to improve SLMs' reasoning abilities using SFT and DPO with remarkable performance, substantially boosting Llama-3 from 52.4% to 71.2% and Mistral from 44.4% to 71.1%. It validates that ReflectEvo can rival or even surpass the reasoning capability of the three prominent open-sourced models on BIG-bench without distillation from superior models or fine-grained human annotation. We further conduct a deeper analysis of the high quality of self-generated reflections and their impact on error localization and correction. Our work highlights the potential of continuously enhancing the reasoning performance of SLMs through iterative reflection learning in the long run.

Improving Meta Introspection of Small LLMs by Learning Self-Reflection from Self-Generated Data

We present GaRAGe, a large RAG benchmark with human-curated long-form answers and annotations of each grounding passage, allowing a fine-grained evaluation of whether LLMs can identify relevant grounding when generating RAG answers. Our benchmark contains 2366 questions of diverse complexity, dynamism, and topics, and includes over 35K annotated passages retrieved from both private document sets and the Web, to reflect real-world RAG use cases. This makes it an ideal test bed to evaluate an LLM’s ability to identify only the relevant information necessary to compose a response, or provide a deflective response when there is insufficient information. Evaluations of multiple state-of-the-art LLMs on GaRAGe show that the models tend to over-summarise rather than (a) ground their answers strictly on the annotated relevant passages (reaching at most a Relevance-Aware Factuality Score of 60%), or (b) deflect when no relevant grounding is available (reaching at most 31% true positive rate in deflections). The F1 in attribution to relevant sources is at most 58.9%, and we show that performance is particularly reduced when answering time-sensitive questions and when having to draw knowledge from sparser private grounding sources.

Downloads

Next from ACL 2025

EuroVerdict: A Multilingual Dataset for Verdict Generation Against Misinformation

Similar lecture

Reconsidering LLM Uncertainty Estimation Methods in the Wild