China

The quality of data plays a vital role in the development of Large-scale Generative Models. Understanding how important a data point is for a generative model is essential for explaining its behavior and improving the performance. The influence function provides a framework for quantifying the impact of individual training data on model predictions. However, the high computational cost has hindered their applicability in large-scale applications. In this work, we present RRInf, a novel and principled method for estimating influence function in large-scale generative AI models. We show that influence function estimation can be transformed into a ridge regression problem. Based on this insight, we develop an algorithm that is efficient and scalable to large models. Experiments on noisy data detection and influential data identification tasks demonstrate that RRInf outperforms existing methods in terms of both efficiency and effectiveness for three commonly used large models: RoBERTa-large, Llama-2-13B-chat, and stable-diffusion-v1.5.

EMNLP 2025

RRInf: Efficient Influence Function Estimation via Ridge Regression for Large Language Models and Text-to-Image Diffusion Models

poster

## 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."<br>

*Message from the General Chair, Dirk Hovy*

[**Link to Conference Handbook**](https://drive.google.com/file/d/1johU5QqVVYO4RfH7QcIORr7qrVBdzdwC/view?usp=sharing)





<br>

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

Sparse AutoEncoders (SAEs) have recently been employed as a promising unsupervised approach for understanding the representations of layers of Large Language Models (LLMs). However, with the growth in model size and complexity, training SAEs is computationally intensive, as typically one SAE is trained for each model layer. To address such limitation, we propose \textit{Group-SAE}, a novel strategy to train SAEs. Our method considers the similarity of the residual stream representations between contiguous layers to group similar layers and train a single SAE per group. To balance the trade-off between efficiency and performance, we further introduce \textit{AMAD} (Average Maximum Angular Distance), an empirical metric that guides the selection of an optimal number of groups based on representational similarity across layers. Experiments on models from the Pythia family show that our approach significantly accelerates training with minimal impact on reconstruction quality and comparable downstream task performance and interpretability over baseline SAEs trained layer by layer. This method provides an efficient and scalable strategy for training SAEs in modern LLMs.

Group-SAE: Efficient Training of Sparse Autoencoders for Large Language Models via Layer Groups

We introduce Debate Speech Evaluation as a novel and challenging benchmark for assessing LLM judges. Evaluating debate speeches requires a deep understanding of the speech at multiple levels, including argument strength and relevance, the coherence and organization of the speech, the appropriateness of its style and tone, and so on. This task involves a unique set of cognitive abilities that previously received limited attention in systematic LLM benchmarking. To explore such skills, we leverage a dataset of over 600 meticulously annotated debate speeches and present the first in-depth analysis of how state-of-the-art LLMs compare to human judges on this task. Our findings reveal a nuanced picture: while larger models can approximate individual human judgments in some respects, they differ substantially in their overall judgment behavior. We also investigate the ability of frontier LLMs to generate persuasive, opinionated speech, showing that models may perform at a human level on this task. We make our code and data available at https://anonymous.4open.science/r/BenchmarkingLLMAJ-220C.

Debatable Intelligence: Benchmarking LLM Judges via Debate Speech Evaluation

Large language models (LLMs) often acquire knowledge during pretraining that is undesirable in downstream deployments, e.g., sensitive information or copyrighted content. Existing approaches for removing such knowledge rely on fine-tuning, training low-rank adapters or fact-level editing, but these are either too coarse, too shallow, or ineffective. In this work, we propose PISCES, a novel framework for precisely erasing entire concepts from model parameters by directly editing directions that encode them in parameter space. PISCES uses a disentangler model to decompose MLP vectors into interpretable features, identifies those associated with a target concept using automated interpretability techniques, and removes them from model parameters. Experiments on Gemma 2 and Llama 3.1 over various concepts show that PISCES achieves modest gains in efficacy over leading erasure methods, reducing accuracy on the target concept to as low as 7.7%, while dramatically improving erasure specificity (by up to 31%) and robustness (by up to 41%). Overall, these results demonstrate that feature-based in-parameter editing enables a more precise and reliable approach for removing conceptual knowledge in language models.

Precise In-Parameter Concept Erasure in Large Language Models

We introduce and study artificial impressions--patterns in LLMs’ internal representations of prompts that resemble human impressions and stereotypes. We fit linear probes on generated prompts to predict impressions according to the two-dimensional Stereotype Content Model (SCM). Using these probes, we study the relationship between impressions and downstream model behavior as well as prompt features that inform such impressions. We find that LLMs inconsistently report impressions when prompted, but also that impressions are more consistently linearly decodable from their hidden representations. Additionally, we show that artificial impressions of prompts are predictive of the quality and use of hedging in model responses. We also investigate how particular content, stylistic, and use of dialect in prompts impact LLM impressions.

Artificial Impressions: Evaluating Large Language Model Behavior Through the Lens of Trait Impressions

Stance detection is a task aimed at identifying and analyzing the author's stance from text. Previous studies have primarily focused on the text, which may not fully capture the implicit stance conveyed by the author. To address this limitation, we propose a novel approach that transforms original texts into artificially generated images and uses the visual representation to enhance stance detection. Our approach first employs a text-to-image model to generate candidate images for each text. These images are carefully crafted to adhere to three specific criteria: textual relevance, target consistency, and stance consistency. Next, we introduce a comprehensive evaluation framework to select the optimal image for each text from its generated candidates. Subsequently, we introduce a multimodal stance detection model that leverages both the original textual content and the generated image to identify the author's stance. Experiments demonstrate the effectiveness of our approach and highlight the importance of artificially generated images for stance detection.

Exploring Artificial Image Generation for Stance Detection

Modern large language models (LLMs) employ diverse logical inference mechanisms for reasoning, making the strategic optimization of these approaches critical for advancing their capabilities. This paper systematically investigate the **comparative dynamics** of inductive (System 1) versus abductive/deductive (System 2) inference in LLMs. We utilize a controlled analogical reasoning environment, varying modality (textual, visual, symbolic), difficulty, and task format (MCQ / free-text). Our analysis reveals System 2 pipelines generally excel, particularly in visual/symbolic modalities and harder tasks, while System 1 is competitive for textual and easier problems. Crucially, task format significantly influences their relative advantage, with System 1 sometimes outperforming System 2 in free-text rule-execution. These core findings generalize to broader in-context learning. Furthermore, we demonstrate that advanced System 2 strategies like hypothesis selection and iterative refinement can substantially scale LLM reasoning. This study offers foundational insights and actionable guidelines for strategically deploying logical inference to enhance LLM reasoning.

LogiDynamics: Unraveling the Dynamics of Inductive, Abductive and Deductive Logical Inferences in LLM Reasoning

Autoregressive neural language models (LMs) generate a probability distribution over tokens at each time step given a prompt. In this work, we attempt to systematically understand the probability distributions that LMs can produce, showing that some distributions are significantly harder to elicit than others. Specifically, for any target next-token distribution over the vocabulary, we attempt to find a prompt that induces the LM to output a distribution as close as possible to the target, using either soft or hard gradient-based prompt tuning. We find that (1) in general, distributions with very low or very high entropy are easier to approximate than those with moderate entropy; (2) among distributions with the same entropy, those containing ''outlier tokens'' are easier to approximate; (3) target distributions generated by LMs -- even LMs with different tokenizers -- are easier to approximate than randomly chosen targets. These results offer insights into the expressiveness of LMs and the challenges of using them as probability distribution proposers.

Distribution Prompting: Understanding the Expressivity of Language Models Through the Next-Token Distributions They Can Produce

With the development and widespread application of large language models (LLMs), the new paradigm of ``Model as Product'' is rapidly evolving, and demands higher capabilities to address complex user needs, often requiring precise workflow execution which involves the accurate understanding of multiple tasks. However, existing benchmarks focusing on single-task environments with limited constraints, lack the complexity required to fully reflect To bridge this gap, we present the Extremely Complex Instruction Following Benchmark (EIFBENCH), meticulously crafted to facilitate a more realistic and robust evaluation of LLMs. EIFBENCH not only includes multi-task scenarios that enable comprehensive assessment across diverse task types concurrently, but also integrates a variety of constraints, replicating complex operational environments. Furthermore, we propose the Segment Policy Optimization (SegPO) algorithm to enhance the LLM's ability to accurately fulfill multi-task workflow. Evaluations on EIFBENCH have unveiled considerable performance discrepancies in existing LLMs when challenged with these extremely complex instructions. This finding underscores the necessity for ongoing optimization to navigate the intricate challenges posed by real-world LLM applications.

EIFBENCH: Extremely Complex Instruction Following Benchmark for Large Language Models

Large language models (LLMs) have made significant breakthroughs in extracting useful information from conversation history to enhance the response in long-term conversations. Summarizing useful information from historical conversations has achieved remarkable performance, which, however, may introduce irrelevant or redundant information, making it difficult to flexibly choose and integrate key information from different sessions during memory retrieval. To address this issue, we propose a Fragment-then-Compose framework, a novel memory utilization approach for long-term open-domain conversation, called *FraCom*. To be specific, inspired by the concept of proposition representation from Cognitive Psychology, we first represent the conversation history as a series of predicates plus arguments for propositional representation to preserve key information useful for memory ("**Fragment**"). Then, we compose propositional graphs for the conversation history based on the connection between shared arguments ("**Compose**"). This essentially allows for flexible and effective utilization of related information in long-term memory for better response generation towards a query. Experimental results on four long-term open-domain conversation datasets demonstrate the effectiveness of our *FraCom* in memory utilization and its ability to enhance response generation for LLMs.

Flexibly Utilize Memory for Long-Term Conversation via a Fragment-then-Compose Framework

Evaluating creativity is challenging, even for humans, not only because of its subjectivity but also because it involves complex cognitive processes. Inspired by work in marketing, we attempt to break down creativity into atypicality and originality. With fine-grained human annotations on these dimensions, we propose a suit of tasks specifically for such a subjective problem. We also evaluate the alignment between state-of-the-art (SoTA) vision language models (VLM) and humans on our proposed benchmark, demonstrating both the promises and challenges of using VLMs for automatic creativity assessment.

Downloads

Next from EMNLP 2025

Group-SAE: Efficient Training of Sparse Autoencoders for Large Language Models via Layer Groups

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES