China

The task of image captioning has recently been gaining popularity, and with it the complex task of evaluating the quality of image captioning models. In this work, we present the first survey and taxonomy of over 70 different image captioning metrics and their usage in hundreds of papers, specifically designed to help users select the most suitable metric for their needs. We find that despite the diversity of proposed metrics, the vast majority of studies rely on only five popular metrics, which we show to be weakly correlated with human ratings. We hypothesize that combining a diverse set of metrics can enhance correlation with human ratings. As an initial step, we demonstrate that a linear regression-based ensemble method, which we call EnsembEval, trained on one human ratings dataset, achieves improved correlation across five additional datasets, showing there is a lot of room for improvement by leveraging a diverse set of metrics.

EMNLP 2025

Surveying the Landscape of Image Captioning Evaluation: A Comprehensive Taxonomy, Trends and Metrics Analysis

survey

image captioning

evaluation

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible. The experience for on-site participants would closely approximate a normal pre-COVID *ACL conference, with 5-6 thematically organized parallel sessions and live Q/A and interactive discussion immediately after the talks. Presentations by virtual participants will be equitably interleaved with those of on-site participants, projected on the auditorium screens as if on-site, and also followed immediately by live Q/A and interactive discussion at a time during reasonable waking hours for the virtual presenter. For all participants, on-site and virtual, who are unable to attend a session due to either time-zone issues or because they are participating in another session live, talk recordings and slides will be available online at a minimum after the live presentation (and in many cases before as well), and questions may be submitted in advance on session-specific discussion boards and answered live in session with the usual visual aids if desired.

<iframe style="width:700px;height:400px" src="https://online.fliphtml5.com/ebtyf/ceby/"  seamless="seamless" scrolling="no" frameborder="0" allowtransparency="true" allowfullscreen="true" ></iframe>

Please Note: The EMNLP registration system is not currently connected to the underline site as we are still in the process of building out EMNLP 2021. You will receive access instructions from underline the week of November 1st. 

Access is given only to EMNLP upon registration, if you have not registered please do so [here](https://2021.emnlp.org/registration).

Registered attendees will receive access the week of November 1st.

EMNLP 2021

EMNLP 2021 is planned to be a hybrid event in Punta Cana, Dominican Republic, with both on-site and fully virtual participation possible.

technical paper

## 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*





<br>

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

A large number of studies rely on closed-style multiple-choice surveys to evaluate cultural alignment in Large Language Models (LLMs). In this work, we challenge this constrained evaluation paradigm and explore more realistic, unconstrained approaches. Using the World Values Survey (WVS) and Hofstede Cultural Dimensions as case studies, we demonstrate that LLMs exhibit stronger cultural alignment in less constrained settings, where responses are not forced. Additionally, we show that even minor changes, such as reordering survey choices, lead to inconsistent outputs, exposing the limitations of closed-style evaluations. Our findings advocate for more robust and flexible evaluation frameworks that focus on specific cultural proxies, encouraging more nuanced and accurate assessments of cultural alignment in LLMs.

Break the Checkbox: Challenging Closed-Style Evaluations of Cultural Alignment in LLMs

When solving NLP tasks with limited labelled data, researchers typically either use a general large language model without further update, or use a small number of labelled samples to tune a specialised smaller model. In this work, we answer an important question -- how many labelled samples are required for the specialised small models to outperform general large models, while taking the performance variance into consideration. By observing the behaviour of fine-tuning, instruction-tuning, prompting and in-context learning on 8 language models, we identify such performance break-even points across 8 representative text classification tasks of varying characteristics. We show that the specialised models often need only few samples (on average 100) to be on par or better than the general ones. At the same time, the number of required labels strongly depends on the dataset or task characteristics, with fine-tuning on binary datasets requiring significantly more samples. When performance variance is taken into consideration, the number of required labels increases on average by 100 - 200%. Finally, larger models do not consistently lead to better performance and lower variance, with 4-bit quantisation having negligible impact.

Comparing Specialised Small and General Large Language Models on Text Classification: 100 Labelled Samples to Achieve Break-Even Performance

Cross-Document Multi-entity question answering (MEQA) demands the integration of scattered information across documents to resolve complex queries involving entities, relationships, and contextual dependencies. Although large language models (LLMs) and retrieval-augmented generation (RAG) systems show promise, their performance on cross-document MEQA remains underexplored due to the absence of tailored benchmarks. To address this gap, we introduce MEBench, a scalable multi-document, multi-entity benchmark designed to systematically evaluate LLMs’ capacity to retrieve, consolidate, and reason over scattered and dense information. Our benchmark comprises 4,780 questions which are systematically categorized into three primary categories: Comparative Reasoning, Statistical Reasoning and Relational Reasoning, further divided into eight distinct types, ensuring broad coverage of real-world multi-entity reasoning scenarios. Our experiments on state-of-the-art LLMs reveal critical limitations: even advanced models achieve only 59% accuracy on MEBench. Our benchmark emphasizes the importance of completeness and factual precision of information extraction in MEQA tasks, using Entity-Attributed F1 (EA-F1) metric for granular evaluation of entity-level correctness and attribution validity. MEBench not only highlights systemic weaknesses in current LLM frameworks but also provides a foundation for advancing robust, entity-aware QA architectures. The source code and data have been made available at https://github.com/tl2309/SRAG.

MEBench: Benchmarking Large Language Models for Cross-Document Multi-Entity Question Answering

Code-mixing, the practice of switching between languages within a conversation, poses unique challenges for traditional NLP. Existing benchmarks like LinCE and GLUECoS are limited by their narrow language pairs and tasks, failing to adequately assess large language models' (LLMs) code-mixing abilities. Despite the recognized importance of code-mixing for multilingual users, research on LLMs in this context remains sparse. Additionally, current techniques for synthesizing code-mixed data are underdeveloped to generate code-mixing. In response, we introduce CodeMixBench, a comprehensive benchmark covering eight tasks, including three specific to LLMs and five traditional NLP tasks, and 18 languages from seven language families. We also propose a new method for generating large-scale synthetic code-mixed texts by combining word substitution with GPT-4 prompting. Our evaluation reveals consistent underperformance of LLMs on code-mixed datasets involving different language families. Enhancements in training data size, model scale, and few-shot learning could improve their performance.

CodeMixBench: Evaluating Code-Mixing Capabilities of LLMs Across 18 Languages

Image Captioning for state-of-the-art VLMs has significantly improved over time; however, this comes at the cost of increased computational complexity, making them less accessible for resource-constrained applications such as mobile devices and assistive technologies. Alternatively, smaller VLMs prioritize high-level scene descriptions, overlooking finer details that contribute to a richer understanding of an image. In this paper, we introduce a training-free framework that enhances caption diversity and informativeness by explicitly attending to distinct image regions using a comparably small VLM, BLIP, as the backbone. Our approach leverages structured segmentation to produce hierarchical representations that capture both global and localized semantics. Without requiring additional model training, we demonstrate that our method allows smaller VLMs to achieve performance comparable to larger models in terms of image-caption alignment, semantic integrity, and diversity. We evaluate our framework on MSCOCO, Flickr30k, and Nocaps test datasets, achieving a Div-2 score of 0.735, 0.750, and 0.748 for each dataset respectively, while maintaining strong image-caption relevancy and semantic integrity with the human-annotated captions.

Image Embedding Sampling Method for Diverse Captioning

Modern BPE tokenisers often split calendar dates into meaningless fragments, e.g., “20250312” rightarrow “202”, “503”, “12”, inflating token counts and obscuring the inherent structure needed for robust temporal reasoning. In this work, we (1) introduce a simple yet interpretable metric, termed date fragmentation ratio, that measures how faithfully a tokeniser preserves multi-digit date components; (2) release \textsc{DateAugBench}, a suite of 6500 examples spanning three temporal reasoning tasks: context-based date resolution, format-invariance puzzles, and date arithmetic across historical, contemporary, and future time periods; and (3) through layer-wise probing and causal attention-hop analyses, uncover an emergent date-abstraction mechanism whereby large language models stitch together the fragments of month, day, and year components for temporal reasoning. Our experiments show that excessive fragmentation correlates with accuracy drops of up to 10 points on uncommon dates like historical and futuristic dates. Further, we find that the larger the model, the faster the emergent date abstraction heals date fragments. Lastly, we observe a reasoning path that LLMs follow to assemble date fragments, typically differing from human interpretation (year rightarrow month rightarrow day).

Date Fragments: A Hidden Bottleneck of Tokenization for Temporal Reasoning

Large Audio-Language Models (LALMs) are augmented with the ability to perceive audio, demonstrating impressive capabilities in processing combined audio and text signals. However, their reliability when faced with conflicting inputs across modalities remains largely unexplored. This study examines how LALMs prioritize information when presented with inconsistent audio-text pairs. Through extensive evaluation across diverse audio understanding tasks, we reveal a concerning phenomenon: when inconsistencies exist between modalities, LALMs display a significant bias toward textual input, often disregarding audio evidence. This tendency leads to substantial performance degradation in audio-centric tasks and raises important reliability concerns for real-world applications. We further investigate the influencing factors of text bias, explore mitigation strategies through supervised fine-tuning, and analyze model confidence patterns that reveal persistent overconfidence even with contradictory inputs. These findings underscore the need for improved modality balancing during training and more sophisticated fusion mechanisms to enhance robustness when handling conflicting multi-modal inputs. The project is available at https://github.com/WangCheng0116/MCR-BENCH.

When Audio and Text Disagree: Revealing Text Bias in Large Audio-Language Models

Model merging has emerged as a promising approach for updating large language models (LLMs) by integrating multiple domain-specific models into a cross-domain merged model. Despite its utility and plug-and-play nature, unmonitored mergers can introduce significant security vulnerabilities, such as backdoor attacks and model merging abuse. In this paper, we identify a novel and more realistic attack surface where a malicious merger can extract targeted personally identifiable information (PII) from an aligned model with model merging. Specifically, we propose Merger-as-a-Stealer, a two-stage framework to achieve this attack: First, the attacker fine-tunes a malicious model to force it to respond to any PII-related queries. The attacker then uploads this malicious model to the model merging conductor and obtains the merged model. Second, the attacker inputs direct PII-related queries to the merged model to extract targeted PII. Extensive experiments demonstrate that Merger-as-a-Stealer successfully executes attacks against various LLMs and model merging methods across diverse settings, highlighting the effectiveness of the proposed framework. Given that this attack enables character-level extraction for targeted PII without requiring any additional knowledge from the attacker, we stress the necessity for improved model alignment and more robust defense mechanisms to mitigate such threats.

Merger-as-a-Stealer: Stealing Targeted PII from Aligned LLMs with Model Merging

Existing Reward Models (RMs), typically trained on general preference data, struggle in Retrieval Augmented Generation (RAG) settings, which require judging responses for faithfulness to retrieved context, relevance to the user query, appropriate refusals when context is insufficient, completeness and conciseness of information. To address the lack of publicly available RAG-centric preference datasets and specialised RMs, we introduce RAGferee, a methodology that repurposes question-answering (QA) datasets into preference pairs that prioritise groundedness over stylistic features, enabling the training of contextual RMs better suited to judging RAG responses. Using RAGferee, we curate a small preference dataset of 4K samples and fine-tune RMs ranging from 7B to 24B parameters. Our RAG-centric RMs achieve state-of-the-art performance on ContextualJudgeBench, surpassing existing 70B+ RMs trained on much larger (up to 2.4M samples) general corpora, with an absolute improvement of +15.5%.

RAGferee: Building Contextual Reward Models for Retrieval-Augmented Generation

Abstract concepts like mercy and peace often lack clear visual grounding, and thus challenge humans and models to provide suitable image representations. To address this challenge, we introduce AbsVis -- a dataset of 675 images annotated with 14,175 concept–explanation attributions from humans and two Vision-Language Models (VLMs: Qwen and LLaVA), where each concept is accompanied by a textual explanation. We compare human and VLM attributions in terms of diversity, abstractness, and alignment, and find that humans attribute more varied concepts. AbsVis also includes 2,680 human preference judgments evaluating the quality of a subset of these annotations, showing that overlapping concepts (attributed by both humans and VLMs) are most preferred. Explanations clarify and strengthen the perceived attributions, both from humans and VLMs. Explanations clarify and strengthen the perceived attributions, both from human and VLMs. Finally, we show that VLMs can approximate human preferences and use them to fine-tune VLMs via Direct Preference Optimization (DPO), yielding improved alignments with preferred concept–explanation pairs.

Downloads

Next from EMNLP 2025

Break the Checkbox: Challenging Closed-Style Evaluations of Cultural Alignment in LLMs

Similar lecture

Retrieval, Analogy, and Composition: A framework for Compositional Generalization in Image Captioning