Strategize Globally, Adapt Locally: A Multi-Turn Red Teaming Agent with Dual-Level Learning

The exploitation of large language models (LLMs) for malicious purposes poses significant security risks as these models become more powerful and widespread. While most existing red-teaming frameworks focus on single-turn attacks, real-world adversaries typically operate in multi-turn scenarios, iteratively probing for vulnerabilities and adapting their prompts based on threat model responses. In this paper, we propose \AlgName, a novel multi-turn red-teaming agent that emulates sophisticated human attackers through complementary learning dimensions: global tactic-wise learning that accumulates knowledge over time and generalizes to new attack goals, and local prompt-wise learning that refines implementations for specific goals when initial attempts fail. Unlike previous multi-turn approaches that rely on fixed strategy sets, \AlgName enables the agent to identify new jailbreak tactics, develop a goal-based tactic selection framework, and refine prompt formulations for selected tactics. Empirical evaluations on JailbreakBench demonstrate our framework's superior performance, achieving over 90\% attack success rates against GPT-3.5-Turbo and Llama-3.1-70B within 5 conversation turns, outperforming state-of-the-art baselines. These results highlight the effectiveness of dynamic learning in identifying and exploiting model vulnerabilities in realistic multi-turn scenarios.

Can LLMs Grasp Implicit Cultural Values? Benchmarking LLMs' Metacognitive Cultural Intelligence with CQ-Bench

Cultural Intelligence (CQ) refers to the ability to understand unfamiliar cultural contexts-a crucial skill for large language models (LLMs) to effectively engage with globally diverse users. While existing research often focuses on explicitly stated cultural norms, such approaches fail to capture the subtle, implicit values that underlie real-world conversations. To address this gap, we introduce CQ-Bench, a benchmark specifically designed to assess LLMs' capability to infer implicit cultural values from natural conversational contexts. We generate a multi-character conversation-based stories dataset using values from the World Value Survey and GlobalOpinions datasets, with topics including ethical, religious, social, and political. Our dataset construction pipeline includes rigorous validation procedures-incorporation, consistency, and implicitness checks-using GPT-4o, with 98.2% human-model agreement in the final validation. Our benchmark consists of three tasks of increasing complexity: attitude detection, value selection, and value extraction. We find that while o1 and Deepseek-R1 models reach human-level performance in value selection (0.809 and 0.814), they still fall short in nuanced attitude detection, with F1 scores of 0.622 and 0.635, respectively. In the value extraction task, GPT-4o-mini and o3-mini score 0.602 and 0.598, highlighting the difficulty of open-ended cultural reasoning. Notably, fine-tuning smaller models (e.g., LLaMA-3.2-3B) on only 500 culturally rich examples improves performance by over 10%, even outperforming stronger baselines (o3-mini) in some cases. Using CQ-Bench, we provide insights into the current challenges in LLMs' CQ research and suggest practical pathways for enhancing LLMs' cross-cultural reasoning abilities.

Discovering Knowledge Deficiencies of Language Models on Massive Knowledge Base

Large language models (LLMs) possess impressive linguistic capabilities but often fail to faithfully retain factual knowledge, leading to hallucinations and unreliable outputs. Understanding LLMs' knowledge deficiencies by exhaustively evaluating against full-scale knowledge bases is computationally prohibitive, especially for closed-weight models. We propose stochastic error ascent (SEA), a scalable and efficient framework for discovering knowledge deficiencies (errors) in closed-weight LLMs under a strict query budget. Rather than naively probing all knowledge candidates, SEA formulates error discovery as a stochastic optimization process: it iteratively retrieves new high-error candidates by leveraging the semantic similarity to previously observed failures. To further enhance search efficiency and coverage, SEA employs hierarchical retrieval across document and paragraph levels, and constructs a relation directed acyclic graph to model error propagation and identify systematic failure modes. Empirically, SEA uncovers 40.7x more knowledge errors than Automated Capability Discovery and 26.7% more than AutoBencher, while reducing the cost-per-error by 599x and 9x, respectively. Human evaluation confirms the high quality of generated questions, while ablation and convergence analyses validate the contribution of each component in SEA. Further analysis on the discovered errors reveals correlated failure patterns across LLM families and recurring deficits, highlighting the need for better data coverage and targeted fine-tuning in future LLM development.

LUME: LLM Unlearning with Multitask Evaluations

Unlearning aims to remove copyrighted, sensitive, or private content from large language models (LLMs) without a full retraining. In this work, we develop a multi-task unlearning benchmark (LUME) which features three tasks: (1) unlearn synthetically generated creative short novels, (2) unlearn synthetic biographies with sensitive information, and (3) unlearn a collection of public biographies. We further release two fine-tuned LLMs of 1B and 7B parameter sizes as the target models. We conduct detailed evaluations of several recently proposed unlearning algorithms and present results on carefully crafted metrics to understand their behavior and limitations.

Attribute Controlled Fine-tuning for Large Language Models: A Case Study on Detoxification

We propose a constraint learning schema for fine-tuning Large Language Models (LLMs) with attribute control. Given a training corpus and control criteria formulated as a sequence-level constraint on model outputs, our method fine-tunes the LLM on the training corpus while enhancing constraint satisfaction with minimal impact on its utility and generation quality. Specifically, our approach regularizes the LLM training by penalizing the KL divergence between the desired output distribution, which satisfies the constraints, and the LLM's posterior. This regularization term can be approximated by an auxiliary model trained to decompose the sequence-level constraints into token-level guidance, allowing the term to be measured by a closed-form formulation. To further improve efficiency, we design a parallel scheme for concurrently updating both the LLM and the auxiliary model. We evaluate the empirical performance of our approach by controlling the toxicity when training an LLM. We show that our approach leads to an LLM that produces fewer inappropriate responses while achieving competitive performance on benchmarks and a toxicity detection task.

Data Advisor: Dynamic Data Curation for Safety Alignment of Large Language Models

Data is a crucial element in large language model (LLM) alignment. Recent studies have explored using LLMs for efficient data collection. However, LLM-generated data often suffers from quality issues, with underrepresented or absent aspects and low-quality datapoints. To address these problems, we propose Data Advisor, an enhanced LLM-based method for generating data that takes into account the characteristics of the desired dataset. Starting from a set of pre-defined principles in hand, Data Advisor monitors the status of the generated data, identifies weaknesses in the current dataset, and advises the next iteration of data generation accordingly. Data Advisor can be easily integrated into existing data generation methods to enhance data quality and coverage. Experiments on safety alignment of three representative LLMs (i.e., Mistral, Llama2, and Falcon) demonstrate the effectiveness of Data Advisor in enhancing model safety against various fine-grained safety issues without sacrificing model utility.

Tree-of-Traversals: A Zero-Shot Reasoning Algorithm for Augmenting Black-box Language Models with Knowledge Graphs

Knowledge graphs (KGs) complement Large Language Models (LLMs) by providing reliable, structured, domain-specific, and up-to-date external knowledge. However, KGs and LLMs are often developed separately and must be integrated after training. We introduce Tree-of-Traversals, a novel zero-shot reasoning algorithm that enables augmentation of black-box LLMs with one or more KGs. The algorithm equips a LLM with actions for interfacing a KG and enables the LLM to perform tree search over possible thoughts and actions to find high confidence reasoning paths. We evaluate on two popular benchmark datasets. Our results show that Tree-of-Traversals significantly improves performance on question answering and KG question answering tasks. Code is available at \url{https://github.com/amazon-science/tree-of-traversals}

Quantitative Certification of Bias in Large Language Models

Large Language Models (LLMs) can produce responses that exhibit social biases and support stereotypes. However, conventional benchmarking is insufficient to thoroughly evaluate LLM bias, as it can not scale to large sets of prompts and provides no guarantees. Therefore, we propose a novel certification framework QuaCer-B (Quantitative Certification of Bias) that provides formal guarantees on obtaining unbiased responses from target LLMs under large sets of prompts. A certificate consists of high-confidence bounds on the probability of obtaining biased responses from the LLM for any set of prompts containing sensitive attributes, sampled from a distribution. We illustrate the bias certification in LLMs for prompts with various prefixes drawn from given distributions. We consider distributions of random token sequences, mixtures of manual jailbreaks, and jailbreaks in the LLM's embedding space to certify its bias. We certify popular LLMs with QuaCer-B and present novel insights into their biases.

TI-ASU: Toward Robust Automatic Speech Understanding through Text-to-speech Imputation Against Missing Speech Modality

Automatic Speech Understanding (ASU) aims at human-like speech interpretation, providing nuanced intent, emotion, sentiment, and content understanding from speech and language (text) content conveyed in speech. Typically, training a robust ASU model relies heavily on acquiring large-scale, high-quality speech and associated transcriptions. However, it is often challenging to collect or use speech data for training ASU due to concerns such as privacy. To approach this setting of enabling ASU when speech (audio) modality is missing, we propose TI-ASU, using a pre-trained text-to-speech model to impute the missing speech. We report extensive experiments evaluating TI-ASU on various missing scales, both multi- and single-modality settings, and the use of LLMs. Our findings show that TI-ASU yields substantial benefits to improve ASU in scenarios where even up to 95% of training speech is missing. Moreover, we show that TI-ASU is adaptive to dropout training, improving model robustness in addressing missing speech during inference.

Toward Informal Language Processing: Knowledge of Slang in Large Language Models

Recent advancement in large language models (LLMs) has offered a strong potential for natural language systems to process informal language. A representative form of informal language is slang, used commonly in daily conversations and online social media. To date, slang has not been comprehensively evaluated in LLMs due partly to the absence of a carefully designed and publicly accessible benchmark. Using movie subtitles, we construct a dataset that supports evaluation on a diverse set of tasks pertaining to automatic processing of slang. For both evaluation and finetuning, we show the effectiveness of our dataset on two core applications: 1) slang detection, and 2) identification of regional and historical sources of slang from natural sentences. We also show how our dataset can be used to probe the output distributions of LLMs for interpretive insights. We find that while LLMs such as GPT-4 achieve good performance in a zero-shot setting, smaller BERT-like models finetuned on our dataset achieve comparable performance. Furthermore, we show that our dataset enables finetuning of LLMs such as GPT-3.5 that achieve substantially better performance than strong zero-shot baselines. Our work offers a comprehensive evaluation and a high-quality benchmark on English slang based on the OpenSubtitles corpus, serving both as a publicly accessible resource and a platform for applying tools for informal language processing.