We introduce LLM Stinger, a novel approach that leverages Large Language Models (LLMs) to automatically generate adversarial suffixes for jailbreak attacks. Unlike traditional methods, which require complex prompt engineering or white-box access, LLM Stinger uses a reinforcement learning (RL) loop to fine-tune an attacker LLM, generating new suffixes based on existing attacks for harmful questions from the HarmBench benchmark. Our method significantly outperforms existing red-teaming approaches (we compared against 15 of the latest methods), achieving a +57.2% improvement in Attack Success Rate (ASR) on LLaMA2-7B-chat and a +50.3% ASR increase on Claude 2, both models known for their extensive safety measures. Additionally, we achieved a 94.97% ASR on GPT-3.5 and 99.4% on Gemma-2B-it, demonstrating the robustness and adaptability of LLM Stinger across open and closed-source models.

Retrieval-augmented generation (RAG) helps address the limitations of parametric knowledge embedded within a language model (LM). In real world settings, retrieved information can vary in complexity, yet most investigations of LM utilisation of context has been limited to synthetic text. We introduce DRUID (Dataset of Retrieved Unreliable, Insufficient and Difficult-to-understand contexts) with real-world queries and contexts manually annotated for stance. The dataset is based on the prototypical task of automated claim verification, for which automated retrieval of real-world evidence is crucial. We compare DRUID to synthetic datasets (CounterFact, ConflictQA) and find that artificial datasets often fail to represent the complexity and diversity of realistically retrieved context. We show that synthetic datasets exaggerate context characteristics rare in real retrieved data, which leads to inflated context utilisation results, as measured by our novel ACU score. Moreover, while previous work has mainly focused on singleton context characteristics to explain context utilisation, correlations between singleton context properties and ACU on DRUID are surprisingly small compared to other properties related to context source. Overall, our work underscores the need for real-world aligned context utilisation studies to represent and improve performance in real-world RAG settings.

Large Language Models (LLMs) have transformed AI but often struggle with tasks that require domain-specific reasoning and logical alignment. Traditional fine-tuning methods do not leverage the vast amount of symbolic domain-knowledge available to us via symbolic reasoning tools (e.g., provers), and are further limited by sparse rewards and unreliable reward models. We introduce Reinforcement Learning via Symbolic Feedback (RLSF), a novel fine-tuning paradigm where symbolic reasoning tools (e.g., solvers, provers, and algebra systems) provide fine-grained feedback to LLMs. RLSF uses poly-sized certificates (e.g., proofs) generated by symbolic tools to identify and correct errors in model outputs, offering token-level guidance without requiring differentiable reasoning systems. This paradigm bridges the gap between symbolic reasoning and LLM fine-tuning, enabling precise alignment with domain-specific constraints while addressing key limitations of traditional reward signals. Via extensive evaluations, we show that our RLSF-based fine-tuning of LLMs outperforms traditional approaches on five different applications (that have some associated logical or domain constraints), namely, program synthesis from natural language pseudo-code to programming language (+31.43% in functional correctness for Google’s CodeGemma-2b compared to supervised fine-tuning, +17.01% in functional correctness compared to GPT-3.5 – 100× larger), three chemistry tasks (+5.5% exact match for molecule generation, +19.4% exact match for forward synthesis, +33.7% exact match for retrosynthesis, using Meta’s Galactica-1.3b, compared to GPT-4 – 1000× larger), and solving the Game of 24 (+25% success rate using Meta’s Llama2-7b compared to traditional methods, and +7% success rate compared to GPT-3.5 – 25× larger). A key takeaway is that fine-tuning via RLSF enables relatively smaller LLMs to significantly outperform closed-source models that are orders of magnitude larger.

Automated frame analysis of political communication is a popular task in computational social science that is used to study how authors select aspects of a topic to frame its reception.So far, such studies have been narrow, in that they use a fixed set of pre-defined frames and focus only on the text, ignoring the visual contexts in which those texts appear.Especially for framing in the news, this leaves out valuable information about editorial choices, which include not just the written article but also accompanying photographs.To overcome such limitations, we present a method for conducting multi-modal, multi-label framing analysis at scale using large (vision-) language models.Grounding our work in framing theory, we extract latent meaning embedded in images used to convey a certain point and contrast that to the text by comparing the respective frames used.We also identify highly partisan framing of topics with issue-specific frame analysis found in prior qualitative work.We demonstrate a method for doing scalable integrative framing analysis of both text and image in news, providing a more complete picture for understanding media bias.

Nadav Borenstein, Arnav Arora, Lucie-Aimée Kaffee, Isabelle Augenstein. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Yong Cao, Haijiang Liu, Arnav Arora, Isabelle Augenstein, Paul Röttger, Daniel Hershcovich. Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2025.

Stance Detection is concerned with identifying the attitudes expressed by an author towards a target of interest. This task spans a variety of domains ranging from social media opinion identification to detecting the stance for a legal claim. However, the framing of the task varies within these domains, in terms of the data collection protocol, the label dictionary and the number of available annotations. Furthermore, these stance annotations are significantly imbalanced on a per-topic and inter-topic basis. These make multi-domain stance detection a challenging task, requiring standardization and domain adaptation. To overcome this challenge, we propose Topic Efficient StancE Detection (TESTED), consisting of a topic-guided diversity sampling technique and a contrastive objective that is used for fine-tuning a stance classifier. We evaluate the method on an existing benchmark of 16 datasets with in-domain, i.e. all topics seen and out-of-domain, i.e. unseen topics, experiments. The results show that our method outperforms the state-of-the-art with an average of 3.5 F1 points increase in-domain, and is more generalizable with an averaged increase of 10.2 F1 on out-of-domain evaluation while using ≤ 10% of the training data. We show that our sampling technique mitigates both inter- and per-topic class imbalances. Finally, our analysis demonstrates that the contrastive learning objective allows the model a more pronounced segmentation of samples with varying labels.

The moderation of content on online platforms is usually non-transparent.On Wikipedia, however, this discussion is carried out publicly and editors are encouraged to use the content moderation policies as explanations for making moderation decisions.Currently, only a few comments explicitly mention those policies -20% of the English ones, but as few as 2% of the German and Turkish comments.To aid in this process of understanding how content is moderated, we construct a novel multilingual dataset of Wikipedia editor discussions along with their reasoning in three languages.The dataset contains the stances of the editors (keep, delete, merge, comment), along with the stated reason, and a content moderation policy, for each edit decision.We demonstrate that stance and corresponding reason (policy) can be predicted jointly with a high degree of accuracy, adding transparency to the decision-making process.We release both our joint prediction models and the multilingual content moderation dataset for further research on automated transparent content moderation.

The goal of stance detection is to determine the viewpoint expressed in a piece of text towards a target. These viewpoints or contexts are often expressed in many different languages depending on the user and the platform, which can be a local news outlet, a social media platform, a news forum, etc. Most research on stance detection, however, has been limited to working with a single language and on a few limited targets, with little work on cross-lingual stance detection. Moreover, non-English sources of labelled data are often scarce and present additional challenges. Recently, large multilingual language models have substantially improved the performance on many non-English tasks, especially such with a limited number of examples. This highlights the importance of model pre-training and its ability to learn from few examples. In this paper, we present the most comprehensive study of cross-lingual stance detection to date: we experiment with 15 diverse datasets in 12 languages from 6 language families, and with 6 low-resource evaluation settings each. For our experiments, we build on pattern-exploiting training (PET), proposing the addition of a novel label encoder to simplify the verbalisation procedure. We further propose sentiment-based generation of stance data for pre-training, which shows sizeable improvement of more than 6% F1 absolute in few-shot learning settings compared to several strong baselines.

Stance detection concerns the classification of a writer's viewpoint towards a target.There are different task variants, e.g., stance of a tweet vs. a full article, or stance with respect to a claim vs. an (implicit) topic.Moreover, task definitions vary, which includes the label inventory, the data collection, and the annotation protocol.All these aspects hinder cross-domain studies, as they require changes to standard domain adaptation approaches.In this paper, we perform an in-depth analysis of 16 stance detection datasets, and we explore the possibility for cross-domain learning from them.Moreover, we propose an end-to-end unsupervised framework for outof-domain prediction of unseen, user-defined labels.In particular, we combine domain adaptation techniques such as mixture of experts and domain-adversarial training with label embeddings, and we demonstrate sizable performance gains over strong baselines, both (i) indomain, i.e., for seen targets, and (ii) out-ofdomain, i.e., for unseen targets.Finally, we perform an exhaustive analysis of the crossdomain results, and we highlight the important factors influencing the model performance.

Recent work has proposed multi-hop models and datasets for studying complex natural language reasoning. One notable task requiring multi-hop reasoning is fact checking, where a set of connected evidence pieces leads to the final verdict of a claim. However, existing datasets either do not provide annotations for gold evidence pages, or the only dataset which does (FEVER) mostly consists of claims which can be fact-checked with simple reasoning and is constructed artificially. Here, we study more complex claim verification of naturally occurring claims with multiple hops over interconnected evidence chunks. We: 1) construct a small annotated dataset, PolitiHop, of evidence sentences for claim verification; 2) compare it to existing multi-hop datasets; and 3) study how to transfer knowledge from more extensive in- and out-of-domain resources to PolitiHop. We find that the task is complex and achieve the best performance with an architecture that specifically models reasoning over evidence pieces in combination with in-domain transfer learning.