Austria

This study introduces Crab, a novel Configurable Role-Playing (RP) LLM with Assessing Benchmark, which consists of Role-Centric Dataset Curation, Persona-Embodying LLM Construction, and Comprehensive Benchmark Creation for RP dialogue generation. Distinct from traditional RP models that employ only several preset roles, Crab enables dynamic configuration of desired roles, thereby enhancing related flexibility and adaptability. To effectively train RP-LLMs, we curated the largest RP training dataset. The dataset provides a detailed role overview for each dialogue, including character profile, conversation scenario, and tagged topic, capturing a broad range of role-based behaviors, emotions, and interactions. We also noticed that current benchmarks lack both proper evaluation standards and methods. Thus, to validate RP-LLMs&#39; effectiveness, we introduced a new benchmark containing an evaluation standard, a test dataset with manual annotations, and a reward model RoleRM designed to automatically assess specific aspects of RP while aligning with human perception. Sufficient experiments reveal that RoleRM significantly outperforms ChatGPT and other evaluation methods in conducting fine-grained evaluations of RP. Also, RP-LLMs powered by Crab demonstrate superior performance across various fine-grained aspects.

ACL 2025

Crab: A Novel Configurable Role-Playing LLM with Assessing Benchmark

roleplaying

llms

evaluation

#### Welcome to the 62nd IEEE International Reliability Physics Symposium (IRPS)! 
**The 2024 IEEE International Reliability Physics Symposium (IRPS) was held in Dallas, Texas from April 14 to 18, 2024.** 

[![](https://assets.underline.io/uploads/markdown_image/1/image/9cc005f94b4ba4a81918eb49e0a51057.png)](https://underline.io/events/446/reception)
[![](https://assets.underline.io/uploads/markdown_image/1/image/acae953ba70ae2a57eed7b0250d6dd7d.png)](https://www.irps.org/)
[![](https://assets.underline.io/uploads/markdown_image/1/image/e1806748f2fc69f825e596872a7afdfd.png)](https://drive.google.com/drive/folders/1SmhpLZD9swYxrfP0TYF6327QTIciy3tV?usp=drive_link)
 

**[IRPS 2024 eGuide](https://drive.google.com/file/d/1DHxkWbafodK_D2HlDi7lwRK7KatgBgBD/view?usp=drive_link)** 
Please save the zip file locally and click the "start.pdf" to view the file. 
 
**IRPS 2024 General Chair Welcome Message**

On behalf of the Management Committee of the IEEE International Reliability Physics Symposium (IRPS) and the IRPS Board of Directors, it gives me great pleasure to extend a warm welcome to all who are planning to attend the IRPS 2024, scheduled to take place from April 14-18, 2024, in Dallas, TX. 
Over its 62-year history, IRPS has solidified its position as the premiere conference for engineers and scientists to present new and original work in the field of microelectronics reliability. Drawing participants from the United States, Europe, Asia, and other parts of the world, IRPS aims to deepen our understanding of reliability of semiconductor devices, integrated circuits, and microelectronic assemblies through an improved understanding of both the physics of failure as well as the application environment. IRPS offers diverse opportunities for attendees to enhance their knowledge and understanding of all aspects of microelectronics reliability. It is also an outstanding chance to meet and network with reliability colleagues from around the world.

The IRPS 2024 technical activities will begin with two days of a total of twenty tutorial sessions covering a broad spectrum of emerging reliability topics beyond CMOS device, circuit and system reliability. These sessions are devoted to both newcomers as well as experts. Three reliability Year in Review (YIR) talks focusing on neuromorphic computing, dielectric breakdown in leading edge devices, and GaN technology will be provided on Monday after the tutorials.

The IRPS 2024 technical program will feature four distinguished keynote talks in the plenary sessions delivered by Prof. Shinichi Takagi from The University of Tokyo and Dr. Su Jin Ahn from Samsung on Tuesday, Dr. Rajeev Malik from IBM on Wednesday, and Dr. Sameer Pendharkar from Texas Instruments on Thursday. The technical program, organized into three parallel tracks, consists of contributed and seventeen invited papers from leaders in industry and academia. Additionally, two invited talks by recipients of paper awards from our sister conferences, European Symposium on Reliability of Electron Device Failure Physics and Analysis (ESREF) 2023 and the International Integrated Reliability Workshop (IIRW) 2023, will be given in the technical program. 

Tuesday evening will host ten workshops, offering an opportunity for in-depth, face-to-face discussions on emerging reliability challenges, including the three IRPS 2024 focus topics. A poster session will be held in the Austin Ranch outside the symposium building on Wednesday evening. Both events are sure to be lively. The IRPS 2024 and the International ESD Workshop (IEW) will be co-hosted with IEW registrants participating in both IRPS and IEW programs.

The IRPS 2024 will adopt a hybrid conference style, allowing virtual access to live oral presentations hosted by the Underline. Real-time streaming of all technical events (excluding Workshops and Posters) will be available in the US Central Timezone, providing virtual attendees with Q&A opportunities. For those virtual attendees who cannot watch the live presentations, the material will be available on demand to all attendees after the symposium. All authors will be presenting in person.

Finally, I would like to extend my extreme gratitude to our patrons, exhibitors, and all participants for their invaluable support, which makes IRPS successful. 

Please enjoy the IRPS 2024!

Koji Eriguchi, General Chair 
of 2024 IRPS Management Committee
 
**About IRPS** 
The IEEE International Reliability Physics Symposium (IRPS) is the premiere conference for engineers and scientists to present new and original work in the area of microelectronics reliability. Drawing participants from the United States, Europe, Asia, and all parts of the world, IRPS seeks to understand the reliability of semiconductor devices, integrated circuits, and microelectronic assemblies through an improved understanding of both the physics of failure as well as the application environment. IRPS provides numerous opportunities for attendees to increase their knowledge and understanding of all aspects of microelectronics reliability, and to meet and network with reliability colleagues from around the world.

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

It looks like you are not registered for this event. 

To access the site please register [**here**](https://www.irps.org/registration-fees). 

Please register!

IRPS 2024 Main Conference

The IEEE International Reliability Physics Symposium (IRPS) is the premiere conference for engineers and scientists to present new and original work in the area of microelectronics reliability. IRPS seeks to understand the reliability of semiconductor devices, integrated circuits, and microelectronic assemblies through an improved understanding of both the physics of failure as well as the application environment. 

This study introduces Crab, a novel Configurable Role-Playing (RP) LLM with Assessing Benchmark, which consists of Role-Centric Dataset Curation, Persona-Embodying LLM Construction, and Comprehensive Benchmark Creation for RP dialogue generation. Distinct from traditional RP models that employ only several preset roles, Crab enables dynamic configuration of desired roles, thereby enhancing related flexibility and adaptability. To effectively train RP-LLMs, we curated the largest RP training dataset. The dataset provides a detailed role overview for each dialogue, including character profile, conversation scenario, and tagged topic, capturing a broad range of role-based behaviors, emotions, and interactions. We also noticed that current benchmarks lack both proper evaluation standards and methods. Thus, to validate RP-LLMs' effectiveness, we introduced a new benchmark containing an evaluation standard, a test dataset with manual annotations, and a reward model RoleRM designed to automatically assess specific aspects of RP while aligning with human perception. Sufficient experiments reveal that RoleRM significantly outperforms ChatGPT and other evaluation methods in conducting fine-grained evaluations of RP. Also, RP-LLMs powered by Crab demonstrate superior performance across various fine-grained aspects.

poster

### 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)

Multimodal Large Language Models (MLLMs) enhance visual tasks by integrating visual representations into large language models (LLMs). The textual modality, inherited from LLMs, enables instruction following and in-context learning, while the visual modality boosts downstream task performance through rich semantic content, spatial information, and grounding capabilities. These modalities work synergistically across various visual tasks. Our research reveals a persistent imbalance between these modalities, with text often dominating output generation during visual instruction tuning, regardless of using full or parameter-efficient fine-tuning (PEFT). We found that re-balancing these modalities can significantly reduce trainable parameters, inspiring further optimization of visual instruction tuning. To this end, we introduce Modality Linear Representation-Steering (MoReS), which re-balances intrinsic modalities by steering visual representations through linear transformations in the visual subspace across each model layer. We validated our approach by developing LLaVA Steering, a suite of models using MoReS. Results show that LLaVA Steering requires, on average, 500 times fewer trainable parameters than LoRA while maintaining comparable performance across three visual benchmarks and eight visual question-answering tasks. Finally, we introduce the LLaVA Steering Factory, a platform that enables rapid customization of MLLMs with a component-based architecture, seamlessly integrating state-of-the-art models and evaluating intrinsic modality imbalance. This open-source project facilitates a deeper understanding of MLLMs within the research community.

LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering

Long Context Language Models (LCLMs) have emerged as a new paradigm to perform Information Retrieval (IR), which enables the direct ingestion and retrieval of information by processing an entire corpus in their single context, showcasing the potential to surpass traditional sparse and dense retrieval methods. However, processing a large number of passages within in-context for retrieval is computationally expensive, and handling their representations during inference further exacerbates the processing time; thus, we aim to make LCLM retrieval more efficient and potentially more effective with passage compression. Specifically, we propose a new compression approach tailored for LCLM retrieval, which is trained to maximize the retrieval performance while minimizing the length of the compressed passages. To accomplish this, we generate the synthetic data, where compressed passages are automatically created and labeled as chosen or rejected according to their retrieval success for a given query, and we train the proposed Compression model for Long context Retrieval (CoLoR) with this data via preference optimization while adding the length regularization loss on top of it to enforce brevity. Through extensive experiments on 9 datasets, we show that CoLoR improves the retrieval performance by 6% while compressing the in-context size by a factor of 1.91. Our code is available at: https://github.com/going-doer/CoLoR.

Efficient Long Context Language Model Retrieval with Compression

Grounded text generation models often produce content that deviates from their source material, requiring user verification to ensure accuracy. Existing attribution methods associate entire sentences with source documents, which can be overwhelming for users seeking to fact-check specific claims. In contrast, existing sub-sentence attribution methods may be more precise but fail to align with users' interests. In light of these limitations, we introduce Localized Attribution Queries (LAQuer), a new task that localizes selected spans of generated output to their corresponding source spans, allowing fine-grained and user-directed attribution. We compare two approaches for the LAQuer task, including prompting large language models (LLMs) and leveraging LLM internal representations. We then explore a modeling framework that extends existing attributed text generation methods to LAQuer. We evaluate this framework across two grounded text generation tasks: Multi-document Summarization (MDS) and Long-form Question Answering (LFQA). Our findings show that LAQuer methods significantly reduce the length of the attributed text. Our contributions include: (1) proposing the LAQuer task to enhance attribution usability, (2) suggesting a modeling framework and benchmarking multiple baselines, and (3) proposing a new evaluation setting to promote future research on localized attribution in content-grounded generation.

LAQuer: Localized Attribution Queries in Content-grounded Generation

Large Language Models (LLMs) have shown impressive reasoning capabilities in well-defined problems with clear solutions, such as mathematics and coding. However, they still struggle with complex real-world scenarios like business negotiations, which require strategic reasoning—an ability to navigate dynamic environments and align long-term goals amidst uncertainty.
Existing methods for strategic reasoning face challenges in adaptability, scalability, and transferring strategies to new contexts.
To address these issues, we propose explicit policy optimization (*EPO*) for strategic reasoning, featuring an LLM that provides strategies in open-ended action space and can be plugged into arbitrary LLM agents to motivate goal-directed behavior.
To improve adaptability and policy transferability, we train the strategic reasoning model via multi-turn reinforcement learning (RL), utilizing process rewards and iterative self-play.
Experiments across social and physical domains demonstrate *EPO*'s ability of long-term goal alignment through enhanced strategic reasoning, achieving state-of-the-art performance on social dialogue and web navigation tasks. Our findings reveal various collaborative reasoning mechanisms emergent in *EPO* and its effectiveness in generating novel strategies, underscoring its potential for strategic reasoning in real-world applications. Code and data are available at [https://github.com/lxqpku/EPO](https://github.com/lxqpku/EPO).

EPO: Explicit Policy Optimization for Strategic Reasoning in LLMs via Reinforcement Learning

Conversational recommender systems (CRSs) enhance recommendation quality by engaging users in multi-turn dialogues, capturing nuanced preferences through natural language interactions. However, these systems often face the false negative issue, where items that a user might like are incorrectly labeled as negative during training, leading to suboptimal recommendations. Expanding the label set through data augmentation presents an intuitive solution but faces the challenge of balancing two key aspects: ensuring semantic relevance and preserving the collaborative information inherent in CRS datasets. To address these issues, we propose a novel data augmentation framework that first leverages an LLM-based semantic retriever to identify diverse and semantically relevant items, which are then filtered by a relevance scorer to remove noisy candidates. Building on this, we introduce a two-stage training strategy balancing semantic relevance and collaborative information. Extensive experiments on two benchmark datasets and user simulators demonstrate significant and consistent performance improvements across various recommenders, highlighting the effectiveness of our approach in advancing CRS performance.

Beyond Single Labels: Improving Conversational Recommendation through LLM-Powered Data Augmentation

Autoregressive models have become the de facto choice for sequence generation tasks, but standard approaches treat digits as independent tokens and apply cross-entropy loss, overlooking the coherent structure of numerical sequences. This paper introduces $\textit{\textbf{N}umerical \textbf{T}oken \textbf{I}ntegrity \textbf{Loss} (NTIL)}$ to address this gap. NTIL operates at two levels: (1) token-level, where it extends the Earth Mover's Distance (EMD) to preserve ordinal relationships between numerical values, and (2) sequence-level, where it penalizes the overall discrepancy between the predicted and actual sequences. This dual approach improves numerical prediction and integrates effectively with LLMs/MLLMs. Extensive experiments show significant performance improvements with NTIL.

Advancing Sequential Numerical Prediction in Autoregressive Models

Comparative Question Answering (CQA) lies at the intersection of Question Answering, Argument Mining, and Summarization. It poses unique challenges due to the inherently subjective nature of many questions and the need to integrate diverse perspectives. Although the CQA task can be addressed using recently emerged instruction-following Large Language Models (LLMs), challenges such as hallucinations in their outputs and the lack of transparent argument provenance remain significant limitations.
To address these challenges, we construct a manually curated dataset comprising arguments annotated with their relevance. These arguments are further used to answer comparative questions, enabling precise traceability and faithfulness. Furthermore, we define explicit criteria for an ``ideal'' comparison and introduce a benchmark for evaluating the outputs of various Retrieval-Augmented Generation (RAG) models with respect to argument relevance. All code and data are publicly released to support further research.

How to Compare Things Properly? A Study on Answering Comparative Questions using Argument Summarization

Language models have shown remarkable capabilities in text generation, but precisely controlling their linguistic style remains challenging. Existing methods either lack fine-grained control, require extensive computation, or introduce significant latency. We propose Style Arithmetic (SA), a novel parameter-space approach that first extracts style-specific representations by analyzing parameter differences between models trained on contrasting styles, then incorporates these representations into a base model with precise control over style intensity. Our experiments show that SA achieves three key capabilities: controllability for precise adjustment of styles, transferability for effective style transfer across tasks, and composability for simultaneous control of multiple style dimensions. Compared to alternative methods, SA offers superior effectiveness while achieving optimal computational efficiency. Our approach opens new possibilities for flexible and efficient style control in language models.

Controllable Style Arithmetic with Language Models

Lexical semantic change detection aims to identify shifts in word meanings over time. While existing methods using embeddings from a diachronic corpus pair estimate the degree of change for target words, they offer limited insight into changes at the level of individual usage instances. To address this, we apply Unbalanced Optimal Transport (UOT) to sets of contextualized word embeddings, capturing semantic change through the excess and deficit in the alignment between usage instances. In particular, we propose Sense Usage Shift (SUS), a measure that quantifies changes in the usage frequency of a word sense at each usage instance. By leveraging SUS, we demonstrate that several challenges in semantic change detection can be addressed in a unified manner, including quantifying instance-level semantic change and word-level tasks such as measuring the magnitude of semantic change and the broadening or narrowing of meaning.

Quantifying Lexical Semantic Shift via Unbalanced Optimal Transport

Large language models (LLMs) have majorly advanced NLP and AI, and next to their ability to perform a wide range of procedural tasks, a major success factor is their internalized factual knowledge. Since (Petroni et al., 2019), analyzing this knowledge has gained attention. However, most approaches investigate one question at a time via modest-sized pre-defined samples, introducing an ``availability bias'' (Tverski and Kahnemann, 1973) that prevents the analysis of knowledge (or beliefs) of LLMs beyond the experimenter's predisposition.

To address this challenge, we propose a novel methodology to comprehensively materialize an LLM's factual knowledge through recursive querying and result consolidation. Our approach is a milestone for LLM research, for the first time providing constructive insights into the scope and structure of LLM knowledge (or beliefs).

As a prototype, we extract a knowledge base (KB) comprising 101 million relational triples for over 2.9 million entities from GPT-4o-mini. We use GPTKB to exemplarily analyze GPT-4o-mini's factual knowledge in terms of scale, accuracy, bias, cutoff and consistency, at the same time. Our resource is accessible at https://gptkb.org.

Downloads

Next from ACL 2025

LLaVA Steering: Visual Instruction Tuning with 500x Fewer Parameters through Modality Linear Representation-Steering

Similar lecture

Understanding the Physical Mechanism of RowPress at the Device-Level in Sub-20 nm DRAM