Austria

Despite the outstanding performance of large language models (LLMs) across various NLP tasks, hallucinations in LLMs--where LLMs generate inaccurate responses--remains as a critical problem as it can be directly connected to a crisis of building safe and reliable LLMs. Uncertainty estimation is primarily used to measure hallucination levels in LLM responses so that correct and incorrect answers can be distinguished clearly. This study proposes an effective uncertainty estimation approach, Clustering-based semantic consistency (Cleanse). Cleanse quantifies the uncertainty with the proportion of the intra-cluster consistency in the total consistency between LLM hidden embeddings which contain adequate semantic information of generations, by employing clustering. The effectiveness of Cleanse for detecting hallucination is validated using four off-the-shelf models, LLaMA-7B, LLaMA-13B, LLaMA2-7B and Mistral-7B and two question-answering benchmarks, SQuAD and CoQA.

ACL 2025

Cleanse: Uncertainty Estimation Approach Using Clustering-based Semantic Consistency in LLMs

workshop paper

### 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)

We introduce CUS-QA, a manually curated, open-ended question-answering dataset specialized in local common knowledge.
The dataset contains 3,904 question-answer pairs across three regions (Czechia, Slovakia, Ukraine) and two modalities (text, visual).
CUS-QA is augmented with manual English translations to all questions and answers, as well as automatically generated baseline answers by a number of prominent open-weight large language models.
While the results using automatic metrics reveal differences between the tested models and languages,
the correlation between these metrics and human judgment is yet to be assessed, and our preliminary experiments suggest only a negligible match.

CUS-QA: Local-Knowledge-Oriented Open-Ended Question Answering Dataset

Evaluating the quality and relevance of textual outputs from Large Language Models (LLMs) remains challenging and resource-intensive.
Existing automated metrics often fail to capture the complexity and variability inherent in LLM-generated outputs.
Moreover, these metrics typically rely on explicit reference standards, limiting their use mostly to domains with objective benchmarks.
This work introduces a novel evaluation framework designed to approximate expert-level assessments of LLM-generated content.
The proposed method employs pairwise comparisons of outputs by multiple LLMs, reducing biases from individual models.
An Elo rating system is used to generate stable and interpretable rankings.
Adjustable agreement thresholds—from full unanimity to majority voting—allow flexible control over evaluation confidence and coverage.
The method’s effectiveness is demonstrated through evaluating competency profiles extracted from scientific abstracts.
Preliminary results show that automatically derived rankings correlate well with expert judgments, significantly reducing the need for extensive human intervention.
By offering a scalable, consistent, and domain-agnostic evaluation layer, the framework supports more efficient and reliable quality assessments of LLM outputs across diverse applications.

(Towards) Scalable Reliable Automated Evaluation with Large Language Models

The power of Large Language Models (LLMs) in user workflows has increased the desire to access such technology in everyday work. While the ability to interact with models provides noticeable benefits, it also presents challenges in terms of how much trust a user should put in the system’s responses. This is especially true for external commercial and proprietary models where there is seldom direct access and only a response from an API is provided. While standard evaluation metrics, such as accuracy, provide starting points, they often may not provide enough information to users in settings where the confidence in a system’s response is important due to downstream or real-world impact, such as in Question & Answering (Q&A) workflows. To support users in assessing how accurate Q&A responses from such black-box LLMs scenarios are, we develop an uncertainty estimation framework that provides users with an analysis using a Dirichlet mixture model accessed from probabilities derived from a zero-shot classification model. We apply our framework to responses on the BoolQ Yes/No questions from GPT models, finding the resulting clusters allow a better quantification of uncertainty, providing a more fine-grained quantification of accuracy and precision across the space of model output while still being computationally practical. We further demonstrate its generalizability and reusability of the uncertainty model by applying it to a small set of Q&A collected from U.S. government websites.

Clustering Zero-Shot Uncertainty Estimations to Assess LLM Response Accuracy for Yes/No Q&A

The LLM-as-a-judge paradigm offers a potential solution to scalability issues in human evaluation of large language models (LLMs), but there are still many open questions about its strengths, weaknesses, and potential biases. This study investigates thirteen models, ranging in size and family, as 'judge models' evaluating answers from nine base and instruction-tuned 'exam-taker models'. We find that only the best (and largest) models show reasonable alignment with humans, though they still differ with up to 5 points from human-assigned scores. Our research highlights the need for alignment metrics beyond percent agreement, as judges with high agreement can still assign vastly different scores. We also find that smaller models and the lexical metric contains can provide a reasonable signal in ranking the exam-taker models. Further error analysis reveals vulnerabilities in judge models, such as sensitivity to prompt complexity and a bias toward leniency. Our findings show that even the best judge models differ from humans in this fairly sterile setup, indicating that caution is warranted when applying judge models in more complex scenarios.

Judging the Judges: Evaluating Alignment and Vulnerabilities in LLMs-as-Judges

The rapid rise of Large Language Models (LLMs) has encouraged the development of detectors for machine-generated text, but their high benchmark scores often do not hold up in real-world scenarios, making these detectors unreliable. This paper reviews datasets with AI-generated content and proposes methods to evaluate their quality and adequacy to train robust detectors capable to distinguish human-written and machine-generated texts.

Are AI Datasets Good Enough? A Survey on Quality of Datasets With Machine-Generated Texts

We can think of Visual Question Answering as a (multimodal) conversation between a human and an AI system. Here, we explore the sensitivity of Vision Language Models (VLMs) through the lens of cooperative principles of conversation proposed by Grice. Specifically, even when Grice's maxims of conversation are flouted, humans typically do not have much difficulty in understanding the conversation even though it requires more cognitive effort. Here, we study if VLMs are capable of handling violations to Grice's maxims in a manner that is similar to humans. Specifically, we add modifiers to human-crafted questions and analyze the response of VLMs to these modifiers. We use three state-of-the-art VLMs in our study, namely, GPT-4o, Claude-3.5-Sonnet and Gemini-1.5-Flash on questions from the VQA v2.0 dataset. Our initial results seem to indicate that the performance of VLMs consistently diminish with the addition of modifiers which indicates our approach as a promising direction to understand the limitations of VLMs.

Analyzing the Sensitivity of Vision Language Models in Visual Question Answering

Large Language Models (LLMs) have demonstrated great potential as evaluators of NLG systems, allowing for high-quality, reference-free, and multi-aspect assessments. However, existing LLM-based metrics suffer from two major drawbacks: reliance on proprietary models to generate training data or perform evaluations, and a lack of fine-grained, explanatory feedback. In this paper, we introduce OpeNLGauge, a fully open-source, reference-free NLG evaluation metric that provides accurate explanations based on error spans. OpeNLGauge is available as a two-stage ensemble of larger open-weight LLMs, or as a small fine-tuned evaluation model, with confirmed generalizability to unseen tasks, domains and aspects. Our extensive meta-evaluation shows that OpeNLGauge achieves competitive correlation with human judgments, outperforming state-of-the-art models on certain tasks while maintaining full reproducibility and providing explanations more than twice as accurate.

OpeNLGauge: An Explainable Metric for NLG Evaluation with Open-Weights LLMs

Large language models (LLMs) are very costly and inefficient to update with new information. To address this limitation, retrieval-augmented generation (RAG) has been proposed as a solution that dynamically incorporates external knowledge during inference, improving factual consistency and reducing hallucinations. Despite its promise, RAG systems face practical challenges-most notably, a strong dependence on the quality of the input query for accurate retrieval. In this paper, we investigate the sensitivity of different components in the RAG pipeline to various types of query perturbations. Our analysis reveals that the performance of commonly used retrievers can degrade significantly even under minor query variations. We study each module in isolation as well as their combined effect in an end-to-end question answering setting, using both general-domain and domain-specific datasets. Additionally, we propose an evaluation framework to systematically assess the query-level robustness of RAG pipelines and offer actionable recommendations for practitioners based on the results of more than 1092 experiments we performed.

Investigating the Robustness of Retrieval-Augmented Generation at the Query Level

Table-to-text generation is a challenging subtask of data-to-text generation, where a natural language generation (NLG) system is generating insights from a given data table. Recent research has built on neural language models (LMs), including large language models (LLMs). However, LLMs were shown to memorize common benchmarks, inflating their true performance. Following prior work on dynamic dataset construction, we developed an approach for obtaining up-to-date benchmarks for table-to-text generation, dubbed FreshTab. This dataset family, based on Wikipedia tables, is not affected by the problems of LLM memorization and benchmark contamination, as the underlying tables are newer than the LLM's knowledge cutoff date. We also introduce basic domain labels for each table, allowing for domain-specific evaluation insights. In our experiments with tables from February-March 2025 collected using FreshTab, we show that all recent LLMs perform worse on average than on a comparable set of tables from the earlier LoTNLG/LogicNLG benchmark.

Sourcing Fresh Resources for Table-to-Text Generation Evaluation

This paper presents a comprehensive evaluation framework for aligning Persian Large Language Models (LLMs) with critical ethical dimensions, including safety, fairness, and social norms. It addresses the gaps in existing LLM evaluation frameworks by adapting them to Persian linguistic and cultural contexts. This benchmark creates three types of Persian-language benchmarks: (i) translated data, (ii) new data generated synthetically, and (iii) new naturally collected data. We translate Anthropic Red Teaming data, AdvBench, HarmBench, and DecodingTrust into Persian. Furthermore, we create ProhibiBench-fa, SafeBench-fa, FairBench-fa, and SocialBench-fa as new datasets to address harmful and prohibited content in indigenous culture. Moreover, we collect extensive dataset as GuardBench-fa to consider Persian cultural norms. By combining these datasets, our work establishes a unified framework for evaluating Persian LLMs, offering a new approach to culturally grounded alignment evaluation. A systematic evaluation of Persian LLMs is performed across the three alignment aspects: safety (avoiding harmful content), fairness (mitigating biases), and social norms (adhering to culturally accepted behaviors). We present a publicly available leaderboard that benchmarks Persian LLMs with respect to safety, fairness, and social norms.

Downloads

Next from ACL 2025

CUS-QA: Local-Knowledge-Oriented Open-Ended Question Answering Dataset

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 2025

CUS-QA: Local-Knowledge-Oriented Open-Ended Question Answering Dataset

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES

Downloads