Austria

Large Language Models (LLMs) are increasingly required to generate text that is both factually accurate and diverse across various open-ended applications. However, current stochastic decoding methods struggle to balance such objectives. We introduce Dynamic Focus Decoding (DFD), a novel plug-and-play stochastic approach that resolves this trade-off without requiring additional data, knowledge, or models. DFD adaptively adjusts the decoding focus based on distributional differences across layers, leveraging the modular and hierarchical nature of factual knowledge within LLMs. This dynamic adjustment improves factuality in knowledge-intensive decoding steps and promotes diversity in less knowledge-reliant steps. DFD can be easily integrated with existing decoding methods, enhancing both factuality and diversity with minimal computational overhead. Extensive experiments across seven datasets demonstrate that DFD significantly improves performance, providing a scalable and efficient solution for open-ended text generation.

ACL 2025

Odysseus Navigates the Sirens&#39; Song: Dynamic Focus Decoding for Factual and Diverse Open-Ended Text Generation

stochastic decoding

open-ended text generation

diversity

large language models

factuality

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

ACL 2024

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Odysseus Navigates the Sirens' Song: Dynamic Focus Decoding for Factual and Diverse Open-Ended Text Generation

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)

Current advanced long-context language models offer great potential for real-world software engineering applications. However, progress in this critical domain remains hampered by a fundamental limitation: the absence of a rigorous evaluation framework for long code understanding. To gap this obstacle, we propose a long code understanding benchmark LongCodeU from four aspects (8 tasks) to evaluate LCLMs' long code understanding ability required for practical applications, including code unit perception, intra-code unit understanding, inter-code unit relation understanding, and long code documentation understanding. We evaluate 9 popular LCLMs on LongCodeU (i.e., 6 general models and 3 code models). Our experimental results reveal key limitations in current LCLMs’ capabilities for long code understanding. Particularly, the performance of LCLMs drops dramatically when the long code length is greater than 32K, falling far short of their claimed 128K to 1M context windows. In the four aspects, inter-code unit relation understanding is the most challenging for LCLMs. Our study provides valuable insights for optimizing LCLMs and driving advancements in software engineering.

Benchmarking Long-Context Language Models on Long Code Understanding

Aligning Large Language Models (LLMs) with human values has attracted increasing attention since it provides clarity, transparency, and the ability to adapt to evolving scenarios. In this paper, we introduce a Controlled Value Vector Activation (ConVA) method that directly aligns the internal values of LLMs by interpreting how a value is encoded in their latent representations and modifies relevant activations to ensure consistent values in LLMs. To ensure an accurate and unbiased interpretation, we propose a context-controlled value vector identification method. To consistently control values without sacrificing model performance, we introduce a gated value vector activation method for effective and minimum degree of value control. Experiments show that our method achieves the highest control success rate across 10 basic values without hurting LLM performance and fluency, and ensures target values even with opposite and potentially malicious input prompts. Source code and data are available at https://github.com/hr-jin/ConVA.

Internal Value Alignment in Large Language Models through Controlled Value Vector Activation

In Reinforcement Learning from Human Feedback (RLHF), it is crucial to learn suitable reward models from human feedback to align large language models (LLMs) with human intentions. However, human feedback can often be noisy, inconsistent, or biased, especially when evaluating complex responses. Such feedback can lead to misaligned reward signals, potentially causing unintended side effects during the RLHF process. To address these challenges, we explore the use of influence functions to measure the impact of human feedback on the performance of reward models. We propose a compute-efficient approximation method that enables the application of influence functions to LLM-based reward models and large-scale preference datasets. Our experiments showcase two key applications of influence functions: (1) detecting common labeler biases in human feedback datasets and (2) guiding labelers in refining their strategies to better align with expert feedback. By quantifying the impact of human feedback, we believe that influence functions can enhance feedback interpretability and contribute to scalable oversight in RLHF, helping labelers provide more accurate and consistent feedback. Source code is available at https://github.com/mintaywon/IF_RLHF.

Understanding Impact of Human Feedback via Influence Functions

Most existing studies on evaluating text-to-image (T2I) models primarily focus on evaluating text-image alignment, image quality, and object composition capabilities, with comparatively fewer studies addressing the evaluation of the factuality of the synthesized images, particularly when the images involve knowledge-intensive concepts. In this work, we present T2I-FactualBench—the largest benchmark to date in terms of the number of concepts and prompts specifically designed to evaluate the factuality of knowledge-intensive concept generation. T2I-FactualBench consists of a three-tiered knowledge-intensive text-to-image generation framework, ranging from the basic memorization of individual knowledge concepts to the more complex composition of multiple knowledge concepts. We further introduce a multi-round visual question answering (VQA)-based evaluation framework to assesses the factuality of three-tiered knowledge-intensive text-to-image generation tasks. Experiments on T2I-FactualBench indicate that current state-of-the-art (SOTA) T2I models still leave significant room for improvement. We release our datasets and code at https://github.com/Safeoffellow/T2I-FactualBench.

T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts

With the rapid advancements in large language models (LLMs), debating tasks, such as argument quality assessment and debate process simulation, have made significant progress. However, existing LLM-based debating systems focus on responding to specific arguments while neglecting objective assessments such as authenticity and logical validity. Furthermore, these systems lack a structured approach to optimize across various dimensions—including evaluation metrics, chain-of-thought (CoT) reasoning, and multi-turn debate refinement—thereby limiting their effectiveness. To address these interconnected challenges, we propose a dual-component framework: (1) InspireScore, a novel evaluation system that establishes a multi-dimensional assessment architecture incorporating four subjective criteria (emotional appeal, argument clarity, argument arrangement, and topic relevance) alongside two objective metrics (fact authenticity and logical validity); and (2) InspireDebate, an optimized debating framework employing a phased optimization approach through CoT reasoning enhancement, multi-dimensional Direct Preference Optimization (DPO), and real-time knowledge grounding via web-based Retrieval Augmented Generation (Web-RAG). Empirical evaluations demonstrate that InspireScore achieves 44% higher correlation with expert judgments compared to existing methods, while InspireDebate shows significant improvements, outperforming baseline models by 57%. Source code is available at https://github.com/fywang12/InspireDebate.

InspireDebate: Multi-Dimensional Subjective-Objective Evaluation-Guided Reasoning and Optimization for Debating

Often, the needs and visual abilities differ between the annotator group and the end user group. Generating detailed diagram descriptions for blind and low-vision (BLV) users is one such challenging domain. Sighted annotators could describe visuals with ease, but existing studies have shown that direct generations by them are costly, bias-prone, and somewhat lacking by BLV standards. In this study, we ask sighted individuals to assess---rather than produce---diagram descriptions generated by vision-language models (VLM) that have been guided with latent supervision via a multi-pass inference. The sighted assessments prove effective and useful to professional educators who are themselves BLV and teach visually impaired learners. We release Sightation, a collection of diagram description datasets spanning 5k diagrams and 137k samples for completion, preference, retrieval, question answering, and reasoning training purposes and demonstrate their fine-tuning potential in various downstream tasks.

Sightation Counts: Leveraging Sighted User Feedback in Building a BLV-aligned Dataset of Diagram Descriptions

Personal travel planning is a challenging task that aims to find a feasible plan that not only satisfies diverse constraints but also meets the demands of the user’s explicit and implicit preferences. In this paper, we study how to integrate the user’s implicit preference into the progress of travel planning. We introduce RealTravel, an augmented version of the TravelPlanner by incorporating real user reviews and point-of-interest metadata from Google Local. Based on RealTravel, we propose Personal Travel Solver (PTS), an integrated system that combines LLMs with numerical solvers to generate travel plans that satisfy both explicit constraints and implicit user preferences. PTS employs a novel architecture that seamlessly connects explicit constraint validation with implicit preference modeling through five specialized modules. The experimental results demonstrate the system's effectiveness, achieving better performance than baseline methods, and improvement in the level of personalization. Our data and code are available at [PersonalTravelSolver](https://github.com/cliftclift/PTS).

Personal Travel Solver: A Preference-Driven LLM-Solver System for Travel Planning

Vision-language models (VLMs) achieve remarkable success in single-image tasks. However, real-world scenarios often involve intricate multi-image inputs, leading to a notable performance decline as models struggle to disentangle critical information scattered across complex visual features. In this work, we propose Focus-Centric Visual Chain, a novel paradigm that enhances VLMs’ perception, comprehension, and reasoning abilities in multi-image scenarios. To facilitate this paradigm, we propose Focus-Centric Data Synthesis, a scalable bottom-up approach for synthesizing high-quality data with elaborate reasoning paths. Through this approach, We construct VISC-150K, a large-scale dataset with reasoning data in the form of Focus-Centric Visual Chain, specifically designed for multi-image tasks. Experimental results on seven multi-image benchmarks demonstrate that our method achieves average performance gains of 3.16% and 2.24% across two distinct model architectures, without compromising the general vision-language capabilities. Our study represents a significant step toward more robust and capable vision-language systems that can handle complex visual scenarios.

Weaving Context Across Images: Improving Vision-Language Models through Focus-Centric Visual Chains

Recently, scaling test-time compute on Large Language Models (LLM) has garnered wide attention. However, there has been limited investigation of how various reasoning prompting strategies perform as scaling. In this paper, we focus on a standard and realistic scaling setting: majority voting. We systematically conduct experiments on 6 LLMs $\times$ 8 prompting strategies $\times$ 6 benchmarks. 
Experiment results consistently show that as the sampling time and computational overhead increase, complicated prompting strategies with superior initial performance gradually fall behind simple Chain-of-Thought.
We analyze this phenomenon and provide theoretical proofs. 
Additionally, we propose a probabilistic method to efficiently predict scaling performance and identify the best prompting strategy under large sampling times, eliminating the need for resource-intensive inference processes in practical applications.
Furthermore, we introduce two ways derived from our theoretical analysis to significantly improve the scaling performance. We hope that our research can promote to re-examine the role of complicated prompting, unleash the potential of simple prompting strategies, and provide new insights for enhancing test-time scaling performance. Code is available at https://github.com/MraDonkey/rethinking_prompting.

Rethinking the Role of Prompting Strategies in LLM Test-Time Scaling: A Perspective of Probability Theory

This paper explores whether enhancing temporal reasoning capabilities in Large Language Models (LLMs) can improve the quality of timeline summarisation, the task of summarising long texts containing sequences of events, such as social media threads. We first introduce NarrativeReason, a novel dataset focused on temporal relationships among sequential events within narratives, distinguishing it from existing temporal reasoning datasets that primarily address pair-wise event relationships. Our approach then combines temporal reasoning with timeline summarisation through a knowledge distillation framework, where we first fine-tune a teacher model on temporal reasoning tasks and then distill this knowledge into a student model while simultaneously training it for the task of timeline summarisation. Experimental results demonstrate that our model achieves superior performance on out-of-domain mental health-related timeline summarisation tasks, which involve long social media threads with repetitions of events and a mix of emotions, highlighting the importance and generalisability of leveraging temporal reasoning to improve timeline summarisation.

Downloads

Next from ACL 2025

Benchmarking Long-Context Language Models on Long Code Understanding

Similar lecture

Dynamic Stochastic Decoding Strategy for Open-Domain Dialogue Generation