United States

We present the first systematic evaluation examining format bias in performance of large language models (LLMs). Our approach distinguishes between two categories of an evaluation metric under format constraints to reliably and accurately assess performance: one measures performance when format constraints are adhered to, while the other evaluates performance regardless of constraint adherence. We then define a metric for measuring the format bias of LLMs and establish effective strategies to reduce it. Subsequently, we present our empirical format bias evaluation spanning four commonly used categories---multiple-choice question-answer, wrapping, list, and mapping---covering 15 widely-used formats. Our evaluation on eight generation tasks uncovers significant format bias across state-of-the-art LLMs. We further discover that improving the format-instruction following capabilities of LLMs across formats potentially reduces format bias. Based on our evaluation findings, we study prompting and fine-tuning with synthesized format data techniques to mitigate format bias. Our methods successfully reduce the variance in ChatGPT&#39;s performance among wrapping formats from 235.33 to 0.71 (%^2)

NAACL 2025

LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs

format following

constraint generation

large language models

### 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**. 

ACL 2025

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

We present the first systematic evaluation examining format bias in performance of large language models (LLMs). Our approach distinguishes between two categories of an evaluation metric under format constraints to reliably and accurately assess performance: one measures performance when format constraints are adhered to, while the other evaluates performance regardless of constraint adherence. We then define a metric for measuring the format bias of LLMs and establish effective strategies to reduce it. Subsequently, we present our empirical format bias evaluation spanning four commonly used categories---multiple-choice question-answer, wrapping, list, and mapping---covering 15 widely-used formats. Our evaluation on eight generation tasks uncovers significant format bias across state-of-the-art LLMs. We further discover that improving the format-instruction following capabilities of LLMs across formats potentially reduces format bias. Based on our evaluation findings, we study prompting and fine-tuning with synthesized format data techniques to mitigate format bias. Our methods successfully reduce the variance in ChatGPT's performance among wrapping formats from 235.33 to 0.71 (%^2)

poster

### Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Welcome to the 2025 meeting of the Nations of the Americas Chapter of the Association for Computational Linguistics! I am proud to help organize the first NAACL conference to carry the new name of our organization, one that emphasizes inclusion for all of the Americas. I am also pleased to welcome you to Albuquerque, New Mexico, a state whose unique blend of cultural influences will make for an excellent backdrop for NAACL 2025, especially with this year’s special theme on NLP in a Multicultural World. 
**[Continue reading...](https://drive.google.com/file/d/1jX-qGhqVSZZCIrAnJaz798pflu5Irrdn/view?usp=sharing)**

*- Colin Cherry, Google, NAACL 2025 General Chair* 

[![](https://assets.underline.io/markdown_image/1/image/b087f8a4dc5816d6e1a6514e59c59ac3.png)](https://drive.google.com/file/d/1T96GzPqObXrMln2BMByCSXTSizjTg69P/view?usp=sharing)

You need to log in with the email address you registered with. Access credentials have been sent to your email. 

Please be sure to check your spam and other email folders if you do not see an email confirmation right away.

Please log in to explore this event.

To access NAACL 2025 event page you are required to register. Please follow [**this link**](https://2025.naacl.org/registration/registration/) to register. Access will depend on your registration type.

Please register!

Welcome to 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics

Text toxicity detection systems exhibit significant biases, producing disproportionate rates of false positives on samples mentioning demographic groups. But what about toxicity detection in speech? To investigate the extent to which text-based biases are mitigated by speech-based systems, we produce a set of high-quality group annotations for the multilingual MuTOX dataset, and then leverage these annotations to systematically compare speech- and text-based toxicity classifiers. Our findings indicate that access to speech data during inference supports reduced bias against group mentions, particularly for ambiguous and disagreement-inducing samples. Our results also suggest that improving classifiers, rather than transcription pipelines, is more helpful for reducing group bias. We publicly release our annotations and provide recommendations for future toxicity dataset construction.

On the Role of Speech Data in Reducing Toxicity Detection Bias

In typical multimodal tasks, such as Visual Question Answering (VQA), adversarial attacks targeting a specific image and question can lead large vision-language models (LVLMs) to provide incorrect answers. However, it is common for a single image to be associated with multiple questions, and LVLMs may still answer other questions correctly even for an adversarial image attacked by a specific question. To address this, we introduce the query-agnostic visual attack (QAVA), which aims to create robust adversarial examples that generate incorrect responses to unspecified and unknown questions. Compared to traditional adversarial attacks focused on specific images and questions, QAVA significantly enhances the effectiveness and efficiency of attacks on images when the question is unknown, achieving performance comparable to attacks on known target questions. Our research broadens the scope of visual adversarial attacks on LVLMs in practical settings, uncovering previously overlooked vulnerabilities, particularly in the context of visual adversarial threats. The code is available at https://github.com/btzyd/qava.

QAVA: Query-Agnostic Visual Attack to Large Vision-Language Models

What ethical concerns, if any, do LLM researchers have? We introduce EthiCon, a corpus of 1,580 ethical concern statements extracted from scientific papers published in the ACL Anthology. We extract ethical concern keywords from the statements and show promising results in automating the concern identification process. Through a survey (N=200), we compare the ethical concerns of the corpus to the concerns listed by the general public and professionals in the field. Finally, we compare our retrieved ethical concerns with existing taxonomies and guidelines pointing to gaps and actionable insights.

Ethical Concern Identification in NLP: A Corpus of ACL Anthology Ethics Statements

Studies of human psychology have demonstrated that people are more motivated to extend empathy to in-group members than out-group members (Cikara et al., 2011). In this study, we investigate how this aspect of intergroup relations in humans is replicated by LLMs in an emotion intensity prediction task. In this task, the LLM is given a short description of an experience a person had that caused them to feel a particular emotion; the LLM is then prompted to predict the intensity of the emotion the person experienced on a numerical scale. By manipulating the group identities assigned to the LLM's persona (the ''perceiver'') and the person in the narrative (the ''experiencer''), we measure how predicted emotion intensities differ between in-group and out-group settings. We observe that LLMs assign higher emotion intensity scores to in-group members than out-group members. This pattern holds across all three types of social groupings we tested: race/ethnicity, nationality, and religion. We perform an in-depth analysis on Llama-3.1-8B, the model which exhibited strongest intergroup bias among those tested.

Language Models Predict Empathy Gaps Between Social In-groups and Out-groups

The field of text privatization often leverages the notion of *Differential Privacy* (DP) to provide formal guarantees in the rewriting or obfuscation of sensitive textual data. A common and nearly ubiquitous form of DP application necessitates the addition of calibrated noise to vector representations of text, either at the data- or model-level, which is governed by the privacy parameter ε. However, noise addition almost undoubtedly leads to considerable utility loss, thereby highlighting one major drawback of DP in NLP. In this work, we introduce a new sentence infilling privatization technique, and we use this method to explore the effect of noise in DP text rewriting. We empirically demonstrate that non-DP privatization techniques excel in utility preservation and can find an acceptable empirical privacy-utility trade-off, yet cannot outperform DP methods in empirical privacy protections. Our results highlight the significant impact of noise in current DP rewriting mechanisms, leading to a discussion of the merits and challenges of DP in NLP as well as the opportunities that non-DP methods present.

On the Impact of Noise in Differentially Private Text Rewriting

In-context learning (ICL) has emerged as a capability of large language models (LLMs), enabling them to adapt to new tasks using provided examples. While ICL has demonstrated its strong effectiveness, there is limited understanding of its vulnerability against potential threats. This paper examines ICL's vulnerability to data poisoning attacks. We introduce ICLPoison, an attacking method specially designed to exploit ICL's unique learning mechanisms by identifying discrete text perturbations that influence LLM hidden states. We propose three representative attack strategies, evaluated across various models and tasks. Our experiments, including those on GPT-4, show that ICL performance can be significantly compromised by these attacks, highlighting the urgent need for improved defense mechanisms to protect LLMs' integrity and reliability.

Data Poisoning for In-context Learning

We study the evolution of opinions inside a population of interacting large language models (LLMs). Every LLM needs to decide how much funding to allocate to an item with three initial possibilities: full, partial, or no funding. We identify biases that drive the exchange of opinions based on the LLM's tendency to find consensus with the other LLM's opinion, display caution when specifying funding, and consider ethical concerns in its opinion. We find these biases are affected by the perceived absence of compelling reasons for opinion change, the perceived willingness to engage in discussion, and the distribution of allocation values. Moreover, tensions among biases can lead to the survival of funding for items with negative connotations. We also find that the final distribution of full, partial, and no funding opinions is more diverse when an LLM freely forms its opinion after an interaction than when its opinion is a multiple-choice selection among the three allocation options. In the latter case, consensus is mostly attained. When agents are aware of past opinions, they seek to maintain consistency with them, changing the opinion dynamics. Our study is performed using Llama 3 and Mistral LLMs.

Biases in Opinion Dynamics in Multi-Agent Systems of Large Language Models: A Case Study on Funding Allocation

Existing studies in backdoor defense have predominantly focused on the training phase, overlooking the critical aspect of testing time defense. This gap becomes pronounced in the context of Large Language Models (LLMs) deployed as Web Services, which typically offer only black-box access, rendering training-time defenses impractical. To bridge this gap, this study critically examines the use of demonstrations as a defense mechanism against backdoor attacks in black-box LLMs. With an identified task, we retrieve task-relevant demonstrations from a clean data pool and integrate them with user queries during testing. Importantly, this approach does not necessitate modifications or tuning of the model, nor does it require insight into the model's internal architecture. The alignment properties inherent in in-context learning play a pivotal role in mitigating the impact of backdoor triggers, effectively recalibrating the behavior of compromised models. Our experimental analysis demonstrates that this method robustly defends against both instance-level and instruction-level backdoor attacks, outperforming existing defense baselines across most evaluation scenarios.

Test-time Backdoor Mitigation for Black-Box Large Language Models with Defensive Demonstrations

Gold'' and ''ground truth'' human-mediated labels have error. This error can escape commonly reported metrics of label quality or obscure questions of accuracy, bias, fairness, and usefulness during model evaluation. This study demonstrates methods for answering such questions even in the context of very low reliabilities from expert humans. We analyze human labels, GPT model ratings, and transformer encoder model ratings of the quality of classroom teaching from two LLM architecture families--encoders and GPT decoders. First, we demonstrate that using standard metrics in the presence of poor labels can mask both label and model quality. The encoder family of models achieve state-of-the-art, even ''super-human'', results across all classroom annotation tasks using standard metrics. However, evaluation techniques accounting for unreliable labels reveal important flaws, including spurious correlations and nonrandom racial biases across models and humans. We estimate that if models were used in a human-in-the-loop context, the variance contributed by GPT model labels would worsen ratings. These techniques also highlight tasks where encoders could offer 80% reduction in human costs while also reducing bias.

All that Glitters'': Techniques for Evaluations with Unreliable Model and Human Annotations

Recent advancements in Large Language Models (LLMs) have significantly enhanced interactions between users and models. These advancements concurrently underscore the need for rigorous safety evaluations due to the manifestation of social biases, which can lead to harmful societal impacts. Despite these concerns, existing benchmarks may overlook the intrinsic weaknesses of LLMs, which can generate biased responses even with simple adversarial instructions. To address this critical gap, we introduce a new benchmark, Fairness Benchmark in LLM under Extreme Scenarios (FLEX), designed to test whether LLMs can sustain fairness even when exposed to prompts constructed to induce bias. To thoroughly evaluate the robustness of LLMs, we integrate prompts that amplify potential biases into the fairness assessment. Comparative experiments between FLEX and existing benchmarks demonstrate that traditional evaluations may underestimate the inherent risks in models. This highlights the need for more stringent LLM evaluation benchmarks to guarantee safety and fairness.

Downloads

Next from NAACL 2025

On the Role of Speech Data in Reducing Toxicity Detection Bias

Similar lecture

From Lists to Emojis: How Format Bias Affects Model Alignment