Austria

Supervised fine-tuning (SFT) enables large language models to align with training data for better performance in many aspects. Nevertheless, the gap between the distribution of current datasets from human annotations or model generations and the real-world data distribution heavily limits the capacities and potentials of models. As a result, we propose a new SFT technique, ATLANTIS, to bridge the gap. We adopt importance sampling to estimate the optimal data distribution in the real world from existing training datasets because the former is hard to sample from. Furthermore, we introduce an extra small model and reference model to estimate the sampling ratio through the probability gap between them. We evaluate our method with benchmarks in knowledge \&amp; understanding and preference aspects. The experiment results prove that ATLANTIS can bring consistent and significant improvements to models&#39; performance. What&#39;s more, our method can be flexibly transferred among models with different structures. Our analyses demonstrate that our method is well-compatible with other SFT techniques to further enhance models&#39; capacities and has great potential to be combined with existing training frameworks.

ACL 2025

ATLANTIS: Weak-to-Strong Learning via Importance Sampling

weak-to-strong learning

supervised fine-tuning

importance sampling

## 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.

Supervised fine-tuning (SFT) enables large language models to align with training data for better performance in many aspects. Nevertheless, the gap between the distribution of current datasets from human annotations or model generations and the real-world data distribution heavily limits the capacities and potentials of models. As a result, we propose a new SFT technique, ATLANTIS, to bridge the gap. We adopt importance sampling to estimate the optimal data distribution in the real world from existing training datasets because the former is hard to sample from. Furthermore, we introduce an extra small model and reference model to estimate the sampling ratio through the probability gap between them. We evaluate our method with benchmarks in knowledge \& understanding and preference aspects. The experiment results prove that ATLANTIS can bring consistent and significant improvements to models' performance. What's more, our method can be flexibly transferred among models with different structures. Our analyses demonstrate that our method is well-compatible with other SFT techniques to further enhance models' capacities and has great potential to be combined with existing training frameworks.

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)

Large vision language models (LVLMs) have improved the document understanding capabilities remarkably, enabling the handling of complex document elements, longer contexts, and a wider range of tasks. However, existing document understanding benchmarks have been limited to handling only a small number of pages and fail to provide a comprehensive analysis of layout elements locating. In this paper, we first define three primary task categories: Long Document Understanding, numerical Reasoning, and cross-element Locating, and then propose a comprehensive benchmark---LongDocURL---integrating above three primary tasks and comprising 20 sub-tasks categorized based on different primary tasks and answer evidences. Furthermore, we develop a semi-automated construction pipeline and collect 2,325 high-quality question-answering pairs, covering more than 33,000 pages of documents, significantly outperforming existing benchmarks. Subsequently, we conduct comprehensive evaluation experiments on both open-source and closed- source models across 26 different configurations, revealing critical performance gaps in this field. The code and data: https://github.com/dengc2023/LongDocURL.

LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating

As the rapid expansion of Machine Learning as a Service (MLaaS) for language models, concerns over the privacy of client inputs during inference or fine-tuning have correspondingly escalated. Recently, solutions have been proposed to safeguard client privacy by obfuscation techniques. However, the solutions incur notable decline in model utility and mainly focus on classification tasks, rendering them impractical for real-world applications. Moreover, recent studies reveal that these obfuscation, if not well designed, is susceptible to embedding inversion attacks (EIAs). In this paper, we devise ObfusLM, a privacy-preserving MLaaS framework for both classification and generation tasks. ObfusLM leverages a model obfuscation module to achieve privacy protection for both classification and generation tasks. Based on $(k, \epsilon)$-anonymity, ObfusLM includes novel obfuscation algorithms to reach provable security against EIAs. Extensive experiments show that ObfusLM outperforms existing works in utility by 10\% with a nearly 80\% resistance rate against EIAs.

ObfusLM: Privacy-preserving Language Model Service against Embedding Inversion Attacks

A popular end-to-end architecture for selective rationalization is the select-then-predict pipeline, comprising a generator to extract highlights fed to a predictor. Such a cooperative system suffers from suboptimal equilibrium minima due to the dominance of one of the two modules, a phenomenon known as interlocking. While several contributions aimed at addressing interlocking, they only mitigate its effect, often by introducing feature-based heuristics, sampling, and ad-hoc regularizations. We present GenSPP, the first interlocking-free architecture for selective rationalization that does not require any learning overhead, as the above-mentioned. GenSPP avoids interlocking by performing disjoint training of the generator and predictor via genetic global search. Experiments on a synthetic and a real-world benchmark show that our model outperforms several state-of-the-art competitors.

Interlocking-free Selective Rationalization Through Genetic-based Learning

Large Language Models (LLMs) have become increasingly capable of handling diverse tasks with the aid of well-crafted prompts and integration of external tools, but as task complexity rises, the workflow involving LLMs can be complicated and thus challenging to implement and maintain. To address this challenge, we propose APPL, A Prompt Programming Language that acts as a bridge between computer programs and LLMs, allowing seamless embedding of prompts into Python functions, and vice versa. 
APPL provides an intuitive and Python-native syntax, an efficient parallelized runtime with asynchronous semantics, and a tracing module supporting effective failure diagnosis and replaying without extra costs. 
We demonstrate that APPL programs are intuitive, concise, and efficient through representative scenarios including Chain-of-Thought with self-consistency (CoT-SC) and ReAct tool-use agent. We further use LLMs to judge the language design between APPL and previous work, where the results indicate that codes written in APPL are more readable and intuitive. Our code, tutorial and documentation are available at 
https://github.com/appl-team/appl.

APPL: A Prompt Programming Language for Harmonious Integration of Programs and Large Language Model Prompts

Topic modeling is a fundamental task in natural language processing, allowing the discovery of latent thematic structures in text corpora. While Large Language Models (LLMs) have demonstrated promising capabilities in topic discovery, their direct application to topic modeling suffers from issues such as incomplete topic coverage, misalignment of topics, and inefficiency. To address these limitations, we propose LLM-ITL, a novel LLM-in-the-loop framework that integrates LLMs with Neural Topic Models (NTMs). In LLM-ITL, global topics and document representations are learned through the NTM. Meanwhile, an LLM refines these topics using an Optimal Transport (OT)-based alignment objective, where the refinement is dynamically adjusted based on the LLM's confidence in suggesting topical words for each set of input words. With the flexibility of being integrated into many existing NTMs, the proposed approach enhances the interpretability of topics while preserving the efficiency of NTMs in learning topics and document representations. Extensive experiments demonstrate that LLM-ITL helps NTMs significantly improve their topic interpretability while maintaining the quality of document representation. Our code and datasets are available at
https://github.com/Xiaohao-Yang/LLM-ITL

Neural Topic Modeling with Large Language Models in the Loop

Despite their impressive ability to generate high-quality and fluent text, generative large language models (LLMs) also produce hallucinations: statements that are misaligned with established world knowledge or provided input context. However, measuring hallucination can be challenging, as having humans verify model generations on-the-fly is both expensive and time-consuming. In this work, we release HALoGEN, a comprehensive hallucination benchmark consisting of: (1) 10,923 prompts for generative models spanning nine domains including programming, scientific attribution, and summarization, and (2) automatic high-precision verifiers for each use case that decompose LLM generations into atomic units, and verify each unit against a high-quality knowledge source. We use this framework to evaluate ~150,000 generations from 14 language models, finding that even the best-performing models are riddled with hallucinations (sometimes up to 86% of generated atomic facts depending on the domain). We further define a novel error classification for LLM hallucinations based on whether they likely stem from incorrect recollection of training data (Type A errors), or incorrect knowledge in training data (Type B errors), or are fabrication (Type C errors). We hope our framework provides a foundation to enable the principled study of why generative models hallucinate, and advances the development of trustworthy large language models.

HALoGEN: Fantastic LLM Hallucinations and Where to Find Them

The Mamba layer offers an efficient selective state-space model (SSM) that is highly effective in modeling multiple domains, including
NLP, long-range sequence processing, and computer vision. Selective SSMs are viewed as dual models, in which one trains in parallel on the entire sequence via an IO-aware parallel scan, and deploys in an autoregressive manner. We add a third view and show that such models can be viewed as attention-driven models. This new perspective enables us to empirically and theoretically compare the underlying mechanisms to that of the attention in transformers and allows us to peer inside the inner workings of the Mamba model with explainability methods. Our code is publicly available.

The Hidden Attention of Mamba Models

In e-commerce, effective product Attribute Mining (AM) is essential for improving product features and aiding consumer decisions. However, current AM methods often focus on extracting attributes from unimodal text, underutilizing multimodal data. In this paper, we propose a novel framework called Multimodal Self-Correction Instruction Tuning (MSIT) to mine new potential attributes from both images and text with Multimodal Large Language Models. The tuning process involves two datasets: Attribute Generation Tuning Data (AGTD) and Chain-of-Thought Tuning Data (CTTD). AGTD is constructed utilizing in-context learning with a small set of seed attributes, aiding the MLLM in accurately extracting attribute-value pairs from multimodal information. To introduce explicit reasoning and improve the extraction in accuracy, we construct CTTD, which incorporates a structured 5-step reasoning process for self-correction. Finally, we employ a 3-stage inference process to filter out redundant attributes and sequentially validate each generated attribute. Comprehensive experimental results on two datasets show that MSIT outperforms state-of-the-art methods. We will release our code and data in the near future.

Open-World Attribute Mining for E-Commerce Products with Multimodal Self-Correction Instruction Tuning

Deep neural network predictions are notoriously difficult to interpret. Feature attribution methods aim to explain these predictions by identifying the contribution of each input feature. Faithfulness, often evaluated using the area over the perturbation curve (AOPC), reflects feature attributions' accuracy in describing the internal mechanisms of deep neural networks. However, many studies rely on AOPC to compare faithfulness across different models, which we show can lead to false conclusions about models' faithfulness. Specifically, we find that AOPC is sensitive to variations in the model, resulting in unreliable cross-model comparisons. Moreover, AOPC scores are difficult to interpret in isolation without knowing the model-specific lower and upper limits. To address these issues, we propose a normalization approach, Normalized AOPC (NAOPC), enabling consistent cross-model evaluations and more meaningful interpretation of individual scores. Our experiments demonstrate that this normalization can radically change AOPC results, questioning the conclusions of earlier studies and offering a more robust framework for assessing feature attribution faithfulness. Our code is available at https://github.com/JoakimEdin/naopc.

Normalized AOPC: Fixing Misleading Faithfulness Metrics for Feature Attributions Explainability

Instruction tuning is crucial for adapting large language models (LLMs) to align with user intentions. Numerous studies emphasize the significance of the quality of instruction tuning (IT) data, revealing a strong correlation between IT data quality and the alignment performance of LLMs. In these studies, the quality of IT data is typically assessed by evaluating the performance of LLMs trained with that data. However, we identified a prevalent issue in such practice: hyperparameters for training models are often selected arbitrarily without adequate justification. We observed significant variations in hyperparameters applied across different studies, even when training the same model with the same data. In this study, we demonstrate the potential problems arising from this practice and emphasize the need for careful consideration in verifying data quality. Through our experiments on the quality of LIMA data and a selected set of 1,000 Alpaca data points, we demonstrate that arbitrary hyperparameter decisions can make any arbitrary conclusion.

Downloads

Next from ACL 2025

LongDocURL: a Comprehensive Multimodal Long Document Benchmark Integrating Understanding, Reasoning, and Locating

Similar lecture

Massive Supervised Fine-tuning Experiments Reveal How Data, Layer, and Training Factors Shape LLM Alignment Quality