Austria

Prompt engineering for large language models is challenging, as even small prompt perturbations or model changes can significantly impact the generated output texts. Existing evaluation methods of LLM outputs, either automated metrics or human evaluation, have limitations, such as providing limited insights or being labor-intensive. We propose Spotlight, a new approach that combines both automation and human analysis. Based on data mining techniques, we automatically distinguish between random (decoding) variations and systematic differences in language model outputs. This process provides token patterns that describe the systematic differences and guide the user in manually analyzing the effects of their prompts and changes in models efficiently. We create three benchmarks to quantitatively test the reliability of token pattern extraction methods and demonstrate that our approach provides new insights into established prompt data. From a human-centric perspective, through demonstration studies and a user study, we show that our token pattern approach helps users understand the systematic differences of language model outputs. We are further able to discover relevant differences caused by prompt and model changes (e.g. related to gender or culture), thus supporting the prompt engineering process and human-centric model behavior research.

ACL 2025

What&#39;s the Difference? Supporting Users in Identifying the Effects of Prompt and Model Changes Through Token Patterns

token patterns

model differences

data mining for nlp

llm evaluation

human-centric

prompt engineering

### 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)

What's the Difference? Supporting Users in Identifying the Effects of Prompt and Model Changes Through Token Patterns

poster

Oracle Bone Script (OBS) is a vital treasure of human civilization, rich in insights from ancient societies. However, the evolution of written language over millennia complicates its decipherment. In this paper, we propose V-Oracle, an innovative framework that utilizes Large Multi-modal Models (LMMs) for interpreting OBS. V-Oracle applies principles of pictographic character formation and frames the task as a visual question-answering (VQA) problem, establishing a multi-step reasoning chain. It proposes a multi-dimensional data augmentation for synthesizing high-quality OBS samples, and also implements a multi-phase oracle alignment tuning to improve LMMs' visual reasoning capabilities. Moreover, to bridge the evaluation gap in the OBS field, we further introduce Oracle-Bench, a comprehensive benchmark that emphasizes process-oriented assessment and incorporates both standard and out-of-distribution setups for realistic evaluation. Extensive experimental results can demonstrate the effectiveness of our method in providing quantitative analyses and superior deciphering capability.

V-Oracle: Making Progressive Reasoning in Deciphering Oracle Bones for You and Me

Mitigating entity bias is a critical challenge in Relation Extraction (RE), where models often rely excessively on entities, resulting in poor generalization. This paper presents a novel approach to address this issue by adapting a Variational Information Bottleneck (VIB) framework. Our method compresses entity-specific information while preserving task-relevant features. It achieves state-of-the-art performance on both general and financial domain RE datasets, excelling in in-domain settings (original test sets) and out-of-domain (modified test sets with type-constrained entity replacements). Our approach offers a robust, interpretable, and theoretically grounded methodology.

A Variational Approach for Mitigating Entity Bias in Relation Extraction

While textless Spoken Language Models (SLMs) have shown potential in end-to-end speech-to-speech modeling, they still lag behind text-based Large Language Models (LLMs) in terms of semantic coherence and relevance. This work introduces the \textbf{Align-SLM} framework, which leverages preference optimization inspired by Reinforcement Learning with Human Feedback (RLHF) to enhance the semantic understanding of SLMs. Our approach generates multiple speech continuations from a given prompt and uses LLM-based semantic metrics to create preference data for Direct Preference Optimization (DPO). We evaluate the framework using ZeroSpeech 2021 benchmarks for lexical and syntactic modeling, the spoken version of the StoryCloze dataset for semantic coherence, and other speech generation metrics, including the GPT4-o score and human evaluation. Experimental results show that our method achieves the state-of-the-art performance of SLMs for most benchmarks, highlighting the importance of preference optimization to improve the semantics of SLMs.

Align-SLM: Textless Spoken Language Models with Reinforcement Learning from AI Feedback

Human perception of events is intrinsically tied to distinguishing between completed (perfect and telic) and ongoing (durative) actions, a process mediated by both linguistic structure and visual cues. In this work, we introduce the Perfect Times dataset, a novel, quadrilingual (English, Italian, Russian, and Japanese) multiple-choice question-answering benchmark designed to assess video-language models (VLMs) on temporal reasoning. By pairing everyday activity videos with event completion labels and perfectivity-tailored distractors, our dataset probes whether models truly comprehend temporal dynamics or merely latch onto superficial markers. Experimental results indicate that state-of-the-art models, despite their success on text-based tasks, struggle to mirror human-like temporal and causal reasoning grounded in video. This study underscores the necessity of integrating deep multimodal cues to capture the nuances of action duration and completion within temporal and causal video dynamics, setting a new standard for evaluating and advancing temporal reasoning in VLMs.

Deep Temporal Reasoning in Video Language Models: A Cross-Linguistic Evaluation of Action Duration and Completion through Perfect Times

Statistical infographics are powerful tools that simplify complex data into visually engaging and easy-to-understand formats. Despite advancements in AI, particularly with LLMs, existing efforts have been limited to generating simple charts, with no prior work addressing the creation of complex infographics from text-heavy documents that demand a deep understanding of the content. We address this gap by introducing the task of generating {\it statistical infographics} composed of multiple sub-charts (e.g., line, bar, pie) that are contextually accurate, insightful, and visually aligned. To achieve this, we define infographic metadata, that includes its title and textual insights, along with sub-chart-specific details such as their corresponding data, alignment, etc. We also present \textbf{\textit{Infodat}}, the first benchmark dataset for text-to-infographic metadata generation, where each sample links a document to its metadata. We propose \textbf{\textit{Infogen}}, a two-stage framework where fine-tuned LLMs first generate metadata, which is then converted into infographic code. Extensive evaluations on \textbf{\textit{Infodat}} demonstrate that \textbf{\textit{Infogen}} achieves state-of-the-art performance, outperforming both closed and open-source LLMs in text-to-statistical infographic generation.

Infogen: Generating Complex Statistical Infographics from Documents

While the embedding of words has revolutionized the field of Natural Language Processing, the embedding of concepts has received much less attention so far. A dense and meaningful representation of concepts, however, could prove useful for several tasks in computational linguistics, especially those involving cross-linguistic data or sparse data from low resource languages. First methods that have been proposed so far embed concepts from automatically constructed colexification networks. While these approaches depart from automatically inferred polysemies, attested across a larger number of languages, they are restricted to the word level, ignoring lexical relations that would only hold for parts of the words in a given language. Building on recently introduced methods for the inference of partial colexifications, we show how they can be used to improve concept embeddings in meaningful ways. The learned embeddings are evaluated against lexical similarity ratings, recorded instances of semantic shift, and word association data. We show that in all evaluation tasks, the inclusion of partial colexifications lead to improved concept representations and better results. Our results further show that the learned embeddings are able to capture and represent different semantic relationships between concepts.

Partial Colexifications Improve Concept Embeddings

We present a new adaptation method MaCP, Minimal yet Mighty adaptive Cosine Projection, that achieves exceptional performance while requiring minimal parameters and memory for fine-tuning large foundation models.
Its general idea is to exploit the superior energy compaction and decorrelation properties of cosine projection to improve both model efficiency and accuracy.
Specifically, it projects the weight change from the low-rank adaptation into the discrete cosine space.
Then, the weight change is partitioned over different levels of the discrete cosine spectrum, and each partition's most critical frequency components are selected.
Extensive experiments demonstrate the effectiveness of MaCP across a wide range of single-modality tasks, including natural language understanding, natural language generation, text summarization, as well as multi-modality tasks such as image classification and video understanding. 
MaCP consistently delivers superior accuracy, significantly reduced computational complexity, and lower memory requirements compared to existing alternatives.

MaCP: Minimal yet Mighty Adaptation via Hierarchical Cosine Projection

Finding evidence for human opinion and behavior at scale is a challenging task, often requiring an understanding of sophisticated thought patterns among vast online communities found on social media. For example, studying how 'gun ownership' is related to the perception of 'Freedom', requires a retrieval system that can operate at scale over social media posts, while dealing with two key challenges: (1) identifying abstract concept instances, (2) which can be instantiated differently across different communities. To address these, we introduce ConceptCarve, an evidence retrieval framework that utilizes traditional retrievers and LLMs to dynamically characterize the search space during retrieval. Our experiments show that ConceptCarve surpasses traditional retrieval systems in finding evidence within a social media community. It also produces an interpretable representation of the evidence for that community, which we use to qualitatively analyze complex thought patterns that manifest differently across the communities.

ConceptCarve: Dynamic Realization of Evidence

Review-based Product Question Answering (PQA) allows e-commerce platforms to automatically address customer queries by leveraging insights from user reviews. However, existing PQA systems generate answers with only a single perspective, failing to capture the diversity of customer opinions. In this paper we introduce a novel task Quantitative Query-Focused Summarization (QQSUM), which aims to summarize diverse customer opinions into representative Key Points (KPs) and quantify their prevalence to effectively answer user queries. While Retrieval-Augmented Generation (RAG) shows promise for PQA, its generated answers still fall short of capturing the full diversity of viewpoints. To tackle this challenge, our model QQSUM-RAG, which extends RAG, employs few-shot learning to jointly train a KP-oriented retriever and a KP summary generator, enabling KP-based summaries that capture diverse and representative opinions. Experimental results demonstrate that QQSUM-RAG achieves superior performance compared to state-of-the-art RAG baselines in both textual quality and quantification accuracy of opinions. Our source code is available at: https://github.com/antangrocket1312/QQSUMM

QQSUM: A Novel Task and Model of Quantitative Query-Focused Summarization for Review-based Product Question Answering

While NLP research has made strides in conversational tasks, many approaches focus on single-turn responses with well-defined objectives or evaluation criteria. In contrast, coaching presents unique challenges with initially undefined goals that evolve through multi-turn interactions, subjective evaluation criteria, mixed-initiative dialogue. In this work, we describe and implement five multi-turn coaching agents that exhibit distinct conversational styles, and evaluate them through a user study, collecting first-person feedback on 155 conversations. We find that users highly value core functionality, and that stylistic components in absence of core components are viewed negatively. By comparing user feedback with third-person evaluations from health experts and an LM, we reveal significant misalignment across evaluation approaches. Our findings provide insights into design and evaluation of conversational coaching agents and contribute toward improving human-centered NLP applications.

Downloads

Next from ACL 2025

V-Oracle: Making Progressive Reasoning in Deciphering Oracle Bones for You and Me

Similar lecture

iPrOp: Interactive Prompt Optimization for Large Language Models with a Human in the Loop