China

Introducing **MARK**, the **M**ulti-st**A**ge **R**easoning framewor**K** for cultural value survey response simulation, designed to enhance the accuracy, steerability, and interpretability of large language models in this task. The system is inspired by the type dynamics theory in the MBTI psychological framework for personality research. It effectively predicts and utilizes human demographic information for simulation: life-situational stress analysis, group-level personality prediction, and self-weighted cognitive imitation. Experiments on the World Values Survey show that MARK outperforms existing baselines by 10% accuracy and reduces the divergence between model predictions and human preferences. This highlights the potential of our framework to improve zero-shot personalization and help social scientists interpret model predictions.

EMNLP 2025

Beyond Demographics: Enhancing Cultural Value Survey Simulation with Multi-Stage Personality-Driven Cognitive Reasoning

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.

The performance of large language models (LLMs) is closely tied to their training data, which can include copyrighted material or private information, raising legal and ethical concerns. Additionally, LLMs face criticism for dataset contamination and internalizing biases. To address these issues, the Pre-Training Data Detection (PDD) task was proposed to identify if specific data was included in an LLM's pre-training corpus. However, existing PDD methods often rely on superficial features like prediction confidence and loss, resulting in mediocre performance. To improve this, we introduce NA-PDD, a novel algorithm analyzing differential neuron activation patterns between training and non-training data in LLMs. This is based on the observation that these data types activate different neurons during LLM inference. We also introduce CCNewsPDD, a temporally unbiased benchmark employing rigorous data transformations to ensure consistent time distributions between training and non-training data. Our experiments demonstrate that NA-PDD significantly outperforms existing methods across three benchmarks and multiple LLMs.

Identifying Pre-training Data in LLMs: A Neuron Activation-Based Detection Framework

Evaluating and iterating upon recommender systems is crucial, yet traditional A/B testing is resource-intensive, and offline methods struggle with dynamic user-platform interactions. While agent-based simulation is promising, existing platforms often lack a mechanism for user actions to dynamically reshape the environment. To bridge this gap, we introduce textbfRecInter, a novel agent-based simulation platform for recommender systems featuring a robust interaction mechanism. In RecInter platform, simulated user actions (e.g., likes, reviews, purchases) dynamically update item attributes in real-time, and introduced textitMerchant Agents can reply, fostering a more realistic and evolving ecosystem. High-fidelity simulation is ensured through textit Multidimensional User Profiling module, textitAdvanced Agent Architecture, and LLM fine-tuned on Chain-of-Thought (CoT) enriched interaction data. Our platform achieves significantly improved simulation credibility and successfully replicates emergent phenomena like textitBrand Loyalty and the textitMatthew Effect. Experiments demonstrate that this interaction mechanism is pivotal for simulating realistic system evolution, establishing our platform as a credible testbed for recommender systems research. All codes are released in https://anonymous.4open.science/r/RecInter-D57E/README.md.

Beyond Static Testbeds: An Interaction-Centric Agent Simulation Platform for Dynamic Recommender Systems

Fine-tuning multi-turn dialogue systems requires high-quality supervision but often suffers from degraded performance when exposed to low-quality data. Supervision errors in early turns can propagate across subsequent turns, undermining coherence and response quality. Existing methods typically address data quality via static prefiltering, which decouples quality control from training and fails to mitigate turn-level error propagation. In this context, we propose **ReSURE** (REgularizing Supervision UnREliability), an adaptive learning method that dynamically down-weights unreliable supervision without explicit filtering. ReSURE estimates per-turn loss distributions using Welford’s online statistics and reweights sample losses on the fly accordingly. Experiments on both single-source and mixed-quality datasets show improved stability and response quality. Notably, ReSURE enjoys positive Spearman correlations (0.21 sim 1.0 across multiple benchmarks) between response scores and number of samples regardless of data quality, which potentially paves the way for utilizing large-scale data effectively.

ReSURE: Regularizing Supervision Unreliability for Multi-turn Dialogue Fine-tuning

The obligatory use of third-person honorifics is a distinctive feature of several South Asian languages, encoding nuanced socio-pragmatic cues such as power, age, gender, fame, and social distance. In this work, (i) We present the first large-scale study of third-person honorific pronoun and verb usage across 10,000 Hindi and Bengali Wikipedia articles with annotations linked to key socio-demographic attributes of the subjects, including gender, age group, fame, and cultural origin. (ii) We uncover systematic intra-language patterns but cross-linguistic inconsistencies: honorifics are more common in Bengali than Hindi, while non-honorifics dominate for infamous, juvenile, and exotic entities. Notably, in both languages, and more prominently in Hindi, men are more frequently addressed with honorifics than women. (iii) To examine whether LLMs internalize similar norms, we probe six LLMs using controlled generation and translation tasks over 1,000 culturally balanced entities. We find that LLMs diverge from Wikipedia trends, exhibiting alternative patterns, raising questions about the real-world validity of their learned patterns, a direction that we leave for future exploration. Our code and data are publicly available at https://anonymous.4open.science/r/honorific-wiki-llm-480F/.

Women, Infamous, and Exotic Beings: A Comparative Study of Honorific Usages in Wikipedia and LLMs for Bengali and Hindi

Connecting audio encoders with large language models (LLMs) allows the LLM to perform various audio understanding tasks, such as automatic speech recognition (ASR) and audio captioning (AC). Most research focuses on training an adapter layer to generate a unified audio feature for the LLM. However, different tasks may require distinct features that emphasize either semantic or acoustic aspects, making task-specific audio features more desirable. In this paper, we propose Prompt-aware Mixture (PaM) to enhance the Speech LLM that uses multiple audio encoders. Our approach involves using different experts to extract different features based on the prompt that indicates different tasks. Experiments demonstrate that with PaM, only one Speech LLM surpasses the best performances achieved by all single-encoder Speech LLMs on ASR, speaker number verification, and AC tasks. PaM also outperforms other feature fusion baselines, such as concatenation and averaging.

Enhancing Speech Large Language Models with Prompt-Aware Mixture of Audio Encoders

Cross-modal retrieval is essential for interpreting cultural heritage data, but its effectiveness is often limited by incomplete or inconsistent textual descriptions, caused by historical data loss and the high cost of expert annotation. While large language models (LLMs) offer a promising solution by enriching textual descriptions, their outputs frequently suffer from hallucinations or miss visually grounded details. To address these challenges, we propose C^3, a data augmentation framework that enhances cross-modal retrieval performance by improving the completeness and consistency of LLM-generated descriptions. C^3 introduces a bidirectional validation mechanism to assess semantic coverage using both visual cues and language-model outputs. Furthermore, to mitigate factual inconsistencies, we formulate a Markov Decision Process to supervise Chain-of-Thought reasoning, guiding consistency verification through adaptive query control. Experiments on the cultural heritage dataset CulTi and general benchmarks MSCOCO and Flickr30K demonstrate that C^3 achieves state-of-the-art performance in both fine-tuned and zero-shot settings.

LLM-Driven Completeness and Consistency Evaluation for Cultural Heritage Data Augmentation in Cross-Modal Retrieval

Speech large language models (LLMs) have emerged as a prominent research focus in speech processing. In this work, we introduce VocalNet, a series of high-performance speech LLMs featuring a scalable and model-agnostic training framework as well as a novel multi-token prediction (MTP) paradigm for speech generation. We first propose an efficient two-stage training framework that enables LLMs to acquire real-time speech interaction capabilities. Through extensive experiments on various training configurations, we ensure both simplicity and effectiveness in the training strategy. Furthermore, inspired by advances in language modeling, we introduce MTP into the domain of speech LLMs—an alternative to traditional next-token prediction (NTP)—which enables the model to predict multiple future tokens at each step. Through systematic analysis and improved implementation, we show that MTP not only accelerates inference speed but also significantly enhances speech quality. Experimental results demonstrate that VocalNet achieves performance comparable to state-of-the-art Omni LLMs while outperforming existing open-source speech LLMs, despite using limited training data.

VocalNet: Speech LLMs with Multi-Token Prediction for Faster and High-Quality Generation

As large language models (LLMs) are increasingly deployed for complex reasoning tasks, Long Chain-of-Thought (Long-CoT) prompting has emerged as a key paradigm for structured inference. Despite early-stage safeguards enabled by alignment techniques such as RLHF, we identify a previously underexplored vulnerability: reasoning trajectories in Long-CoT models can drift from aligned paths, resulting in content that violates safety constraints. We term this phenomenon Path Drift. Through empirical analysis, we uncover three behavioral triggers of Path Drift: (1) first-person commitments that induce goal-driven reasoning that delays refusal signals; (2) ethical evaporation, where surface-level disclaimers bypass alignment checkpoints; (3) condition chain escalation, where layered cues progressively steer models toward unsafe completions. Building on these insights, we introduce a three-stage Path Drift Induction Framework comprising cognitive load amplification, self-role priming, and condition chain hijacking. Each stage independently reduces refusal rates, while their combination further compounds the effect. To mitigate these risks, we propose a path-level defense strategy incorporating role attribution correction and metacognitive reflection (reflective safety cues). Our findings highlight the need for trajectory-level alignment oversight in long-form reasoning beyond token-level alignment.

Path Drift in Large Reasoning Models: How First-Person Commitments Override Safety

Multimodal large language models (MLLMs) have demonstrated extraordinary capabilities in conducting conversations based on image inputs. However, we observe that MLLMs exhibit a pronounced form of visual sycophantic behavior. While similar behavior has also been noted in text-based large language models (LLMs), it becomes significantly more prominent when MLLMs process image inputs. We refer to this phenomenon as the "sycophantic modality gap." To better understand this issue, we further analyze the factors that contribute to the exacerbation of this gap. To mitigate the visual sycophantic behavior, we first experiment with naive supervised fine-tuning to help the MLLM resist misleading instructions from the user. However, we find that this approach also makes the MLLM overly resistant to corrective instructions (i.e., stubborn even if it is wrong). To alleviate this trade-off, we propose Sycophantic Reflective Tuning (SRT), which enables the MLLM to engage in reflective reasoning, allowing it to determine whether a user's instruction is misleading or corrective before drawing a conclusion. After applying SRT, we observe a significant reduction in sycophantic behavior toward misleading instructions, without resulting in excessive stubbornness when receiving corrective instructions.

Pointing to a Llama and Call it a Camel: On the Sycophancy of Multimodal Large Language Models

Subjectivity in NLP tasks, _e.g._, toxicity classification, has emerged as a critical challenge precipitated by the increased deployment of NLP systems in content-sensitive domains. Conventional approaches aggregate annotator judgements (labels), ignoring minority perspectives, and overlooking the influence of the sociocultural context behind such annotations. We propose a framework where subjectivity in binary labels is modeled as an empirical distribution accounting for the variation in annotators through human values extracted from sociocultural descriptors using a language model. The framework also allows for downstream tasks such as population and sociocultural group-level majority label prediction. Experiments on three toxicity datasets covering human-chatbot conversations and social media posts annotated with diverse annotator pools demonstrate that our approach yields well-calibrated toxicity distribution predictions across binary toxicity labels, which are further used for majority label prediction across cultural subgroups, improving over existing methods.

Downloads

Next from EMNLP 2025

Identifying Pre-training Data in LLMs: A Neuron Activation-Based Detection Framework

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES