Austria

Large Language Models (LLMs) have shown remarkable capabilities across various tasks, but their deployment in high-stake domains requires consistent performance across multiple interaction rounds. This paper introduces a comprehensive framework for evaluating and improving LLM response consistency, making three key contributions. First, we propose a novel Position-Weighted Consistency (PWC) score that captures both the importance of early-stage stability and recovery patterns in multi-turn interactions. Second, we present a carefully curated benchmark dataset spanning diverse domains and difficulty levels, specifically designed to evaluate LLM consistency under various challenging follow-up scenarios. Third, we introduce Confidence-Aware Response Generation (CARG), a framework that significantly improves response stability by incorporating model confidence signals into the generation process. Empirical results demonstrate that CARG significantly improves response stability without sacrificing accuracy, underscoring its potential for reliable LLM deployment in critical applications.

ACL 2025

Firm or Fickle? Evaluating Large Language Models Consistency in Sequential Interactions

multi-turn conversations

consistency

large language models

evaluation

## Welcome!
"I am excited to welcome you to this year’s edition of the Conference on Empirical Methods in Natural Language Processing! Importantly, it marks the 30th edition of EMNLP. With over 8,000 submissions, more than 3,000 accepted papers, and thousands of attendees, we have come a long way from that first
workshop, which had 14 accepted papers. As the field looks ahead, Suzhou is the fitting location for celebrating this milestone: rooted in a long literary tradition, yet modern and forward-looking, and home to a large share of the NLP community." 

*Message from the General Chair, Dirk Hovy*

EMNLP 2025

Celebrate 30 Years of EMNLP! 
EMNLP 2025 will be held in Suzhou, China from November 5th to November 9th, 2025.

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)

Recent advancements in large language models (LLMs) have highlighted the importance of improving their reasoning capabilities. 
A critical challenge lies in the scarcity of high-quality reasoning data—characterized by diversity and rich supervisory signals—necessary for robust model training. 
While data augmentation (DA) methods have been leveraged to mitigate this scarcity, prevailing approaches often introduce noise and exhibit logical inconsistencies, thereby diminishing their utility for complex reasoning tasks. 
Moreover, existing DA paradigms predominantly isolate data synthesis from label validation, failing to unify these complementary processes within a cohesive architecture.
To address these limitations, we introduce Logical DA, a multi-agent framework for enhancing reasoning-focused data augmentation in few-shot learning scenarios. 
Our system includes four agents operating through two synergistic phases: (1) diverse data generation, and (2) label verification.
The system incorporates a reflection mechanism to continuously improve data quality by leveraging feedback from logical validation. 
We demonstrate the effectiveness of Logical DA through experiments on various tasks and datasets, achieving the highest average improvement in task accuracy in both fine-tuning and in-context learning paradigms, with an average improvement of 7.61\% when applied to fine-tuning.

Logical DA: Enhancing Data Augmentation for Logical Reasoning via a Multi-Agent System

We introduce KodCode, a synthetic dataset that addresses the persistent challenge of acquiring high-quality, verifiable training data across diverse difficulties and domains for training Large Language Models for coding. Existing code-focused resources typically fail to ensure either the breadth of coverage (e.g., spanning simple coding tasks to advanced algorithmic problems) or verifiable correctness (e.g., unit tests). In contrast, KodCode comprises question–solution–test triplets that are systematically validated via a self-verification procedure. Our pipeline begins by synthesizing a broad range of coding questions, then generates solutions and test cases with additional attempts allocated to challenging problems. Finally, post-training data synthesis is done by rewriting questions into diverse formats and generating responses under a test-based reject sampling procedure from a reasoning model (DeepSeek R1). This pipeline yields a large-scale, robust and diverse coding dataset. It is suitable for supervised fine-tuning and the paired unit tests also provide great potential for RL tuning. Fine-tuning experiments on coding benchmarks (HumanEval(+), MBPP(+), BigCodeBench, and LiveCodeBench) demonstrate that KodCode-tuned models achieve state-of-the-art performance, surpassing models like Qwen2.5-Coder-32B-Instruct and DeepSeek-R1-Distill-Llama-70B.

KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding

Existing Large Language Models (LLMs) enforce uniform computation across all tokens. We analyze the correlation between the input-output difference of self-attention block and Feed-Forward Network (FFN) within the same transformer layer, and find that these two differential vectors are highly correlated. Thus, we propose to dynamically skip the FFN blocks based on the self-attention difference and introduce Diffential Layer Skipping (DiffSkip) to show that LLMs are inherently dynamic-depth models, capable of adjusting computational depth when generating different tokens. DiffSkip employs a lightweight router module to dynamically skip a set of FFN blocks in LLMs and only requires efficient fine-tuning while keeping the whole LLM frozen. Experimental results demonstrate that DiffSkip effectively enables dynamic FFN skipping in decoder-only language models, even in continuous token generation tasks where many layer-skipping methods struggle.

DiffSkip: Differential Layer Skipping in Large Language Models

Recent advancements in large language models (LLMs) have markedly improved their capacity to handle long text inputs; however, current models, including GPT-4o, still exhibit unsatisfactory performance in long-form generation. Generating high-quality long-form content still remains a significant challenge. In this paper, we present LongDPO, a novel approach designed to enhance long-form text generation through step-level supervision. By leveraging Monte Carlo Tree Search (MCTS) to collect stepwise preference pairs and employing a global memory pool to maintain factual accuracy, LongDPO effectively mitigates issues such as inconsistencies that are prevalent in long-context LLMs. Furthermore, we integrate critique-augmented generation to refine the selected preference pairs. Following the collection of stepwise preference pairs, we apply stepwise preference learning for fine-grained optimization. Experimental results demonstrate that our method enhances performance on long-form generation benchmarks (e.g. LongBench-Write) while maintaining nearly lossless performance on several general benchmarks.

LongDPO: Unlock Better Long-form Generation Abilities for LLMs via Critique-augmented Stepwise Information

Code generation is a critical reasoning task for large language models (LLMs). Recent advancements have focused on optimizing the thought process of code generation, achieving significant improvements. However, such thought process lacks effective process supervision, making it hard to optimize the thoughts. Although Process Reward Models (PRMs) have been widely established in mathematical reasoning, building a code PRM is still not trivial for the gap between thoughts to code. In this paper, we propose CodePRM, a novel approach that leverages the code execution feedback to build a code PRM. Specifically, we first collect a large dataset of thought traces, where each thought step is labeled with their derived code' pass rates, accompanied by the corresponding code snippets, and execution feedback. During training, we train a PRM to take both the reasoning process and code execution feedback as input to score individual thought steps, enabling it to leverage code execution results to distinguish between high-quality and low-quality thought steps. Finally, to use the PRM during inference, we develop a Generate-Verify-Refine (GVR) pipeline where the CodePRM serves as a process verifier to dynamically identify and correct errors in the thought process during code search. Experimental results demonstrate that CodePRM with the inference algorithm outperforms strong baselines, significantly enhancing code generation performance. Further analysis reveals the key factors for building a code PRM.

CodePRM: Execution Feedback-enhanced Process Reward Model for Code Generation

Large Language Models (LLMs) perform exceedingly well in Natural Language Understanding (NLU) tasks for many languages including English. However, despite being the ﬁfth most-spoken language globally, Grammatical Error Correction (GEC) in Bangla remains underdeveloped. In this work, we investigate how LLMs can be leveraged for improving Bangla GEC. For that, we ﬁrst do an extensive categorization of 12 error classes in Bangla, and take a survey of native Bangla speakers to collect real-world errors. We next devise a rule-based noise injection method to create grammatically incorrect sentences corresponding to correct ones. The Vaiyākaraṇa dataset, thus created, consists of 567,422 sentences of which 227,119 are erroneous. This dataset is then used to instruction-tune LLMs for the task of GEC in Bangla. Evaluations show that instruction-tuning with Vaiyākaraṇa improves GEC performance of LLMs by 3-7 percentage points as compared to zero-shot setting, and makes them achieve human-like performance in error identiﬁcation. Humans, though, remain superior in grammatical error correction.

Leveraging LLMs for Bangla Grammar Error Correction: Error Categorization, Synthetic Data, and Model Evaluation

While large language models (LLMs) have significantly advanced mathematical reasoning, Process Reward Models (PRMs) have been developed to evaluate the logical validity of reasoning steps. However, PRMs still struggle with out-of-distribution (OOD) challenges. This paper identifies the OOD issues including step OOD, arising from differences in reasoning patterns across model types and sizes, and question OOD, due to dataset shifts between training and real-world problems. To address these issues, we introduce Retrieval-Augmented Process Reward Model (RetrievalPRM), a novel framework designed to tackle these OOD issues. By utilizing a two-stage retrieval-enhanced mechanism, RetrievalPRM retrieves semantically similar questions and steps for PRM as a warmup to stimulate its potential to judge target steps, improving generalization and reasoning consistency across different models and problem types. Our extensive experiments demonstrate that RetrievalPRM outperforms existing baselines across multiple real-world datasets. Our open-source contributions include a retrieval-enhanced dataset, a tuning framework for PRM training, and the RetreivalPRM model, establishing a new standard for PRM performance.

Retrieval-Augmented Process Reward Model for Generalizable Mathematical Reasoning

The latent knowledge of large language models (LLMs) contains harmful or unethical content, which introduces significant security risks upon their widespread deployment. Conducting jailbreak attacks on LLMs can proactively identify vulnerabilities to enhance their security measures. However, previous jailbreak attacks primarily focus on single-turn dialogue scenarios, leaving vulnerabilities in multi-turn dialogue contexts inadequately explored. This paper investigates the resilience of black-box LLMs in multi-turn jailbreak attack scenarios from a novel interrogation perspective. We propose an optimal interrogation principle to conceal the jailbreak intent and introduce a multi-turn attack chain generation strategy called CoA. By employing two effective interrogation strategies tailored for LLMs, coupled with an interrogation history record management mechanis, it achieves a significant optimization of the attack process. Our approach enables the iterative generation of attack chains, offering a powerful tool for LLM red team testing. Experimental results demonstrate that LLMs exhibit insufficient resistance under multi-turn interrogation, with our method shows more advantages(ASR, 83% vs 64%). This work offers new insights into improving the safety of LLMs.

Chain of Attack: Hide Your Intention through Multi-Turn Interrogation

The rapid adoption of large language models (LLMs) in diverse applications has intensified concerns over their security and integrity, especially in cloud environments where internal model parameters are inaccessible to users. Traditional tamper detection methods, designed for deterministic classification models, fail to address the output randomness and massive parameter spaces characteristic of LLMs. In this paper, we introduce \emph{Efficient Sensitive Fingerprinting (ESF)}, the first fingerprinting method tailored for black-box tamper detection of LLMs. ESF generates fingerprint samples by optimizing output sensitivity at selected detection token positions and leverages \emph{Randomness-Set Consistency Checking (RSCC)} to accommodate inherent output randomness. Furthermore, a novel \emph{Max Coverage Strategy (MCS)} is proposed to select an optimal set of fingerprint samples that maximizes joint sensitivity to tampering. Grounded in a rigorous theoretical framework, ESF is both computationally efficient and scalable to large models. Extensive experiments across state-of-the-art LLMs demonstrate that ESF reliably detects tampering, such as fine-tuning, model compression, and backdoor injection, with a detection rate exceeding 99.2\%, thereby offering a robust solution for securing cloud-based AI systems.

ESF: Efficient Sensitive Fingerprinting for Black-Box Tamper Detection of Large Language Models

Solving expert-level multimodal tasks is a key milestone in general intelligence. As the capabilities of multimodal large language models (MLLMs) continue to evolve, evaluation of frontier multimodal intelligence becomes necessary yet challenging. In this work, we introduce ProBench, a benchmark of open-ended user queries encapsulating professional expertise and advanced reasoning. ProBench consists of 4,000 high-quality samples independently collected from professionals based on their productivity demands. It spans across 10 fields and 56 sub-fields, including science, arts, humanities, coding, mathematics, and creative writing. Experimentally, we evaluate and compare 24 latest models using MLLM-as-a-Judge. Our results reveal that although the best open-source models rival the proprietary ones, they all face significant challenges in visual perception, textual understanding, domain knowledge, and advanced reasoning. Our benchmark is publicly accessible at \url{TBC}.

Downloads

Next from ACL 2025

Logical DA: Enhancing Data Augmentation for Logical Reasoning via a Multi-Agent System

Similar lecture

Estimating LLM Consistency: A User Baseline vs Surrogate Metrics