The Multilingual Alignment Prism: Aligning Global and Local Preferences to Reduce Harm

• URL: http://arxiv.org/abs/2406.18682v2
• Date: Mon, 8 Jul 2024 14:26:16 GMT
• Title: The Multilingual Alignment Prism: Aligning Global and Local Preferences to Reduce Harm
• Abstract summary: AI systems are increasingly used across the world, yet safety alignment is often focused on homogeneous monolingual settings. We collect the first set of human annotated red-teaming prompts in different languages distinguishing between global and local harm. We establish a new precedent for state-of-the-art alignment techniques across 6 languages with minimal degradation in general performance.
• Score: 17.566947609534356
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: A key concern with the concept of "alignment" is the implicit question of "alignment to what?". AI systems are increasingly used across the world, yet safety alignment is often focused on homogeneous monolingual settings. Additionally, preference training and safety measures often overfit to harms common in Western-centric datasets. Here, we explore the viability of different alignment approaches when balancing dual objectives: addressing and optimizing for a non-homogeneous set of languages and cultural preferences while minimizing both global and local harms. We collect the first set of human annotated red-teaming prompts in different languages distinguishing between global and local harm, which serve as a laboratory for understanding the reliability of alignment techniques when faced with preference distributions that are non-stationary across geographies and languages. While this setting is seldom covered by the literature to date, which primarily centers on English harm mitigation, it captures real-world interactions with AI systems around the world. We establish a new precedent for state-of-the-art alignment techniques across 6 languages with minimal degradation in general performance. Our work provides important insights into cross-lingual transfer and novel optimization approaches to safeguard AI systems designed to serve global populations.

Related papers

• Inference-Time Steering for Cross-Lingual Factual Consistency in LLMs [51.56484100374058]
  Large Language Models (LLMs) demonstrate remarkable multilingual fluency, but their internal knowledge representations remain disproportionately biased toward high-resource languages.<n>We investigate whether these biases can be mitigated at inference time, forcing an English-prompted model to answer as if it were queried in target languages.<n>Experiments on Gemma 3 12B Instruct reveal persona prompting to be the strongest overall intervention, balancing efficacy, safety, and out-of-domain generalization.
  arXiv  Detail & Related papers  (2026-07-21T16:15:05Z)
• Pluralis v0.1: Towards a Multicultural, Multimodal, Multilingual Benchmark for AI Risk and Reliability [46.80780735011334]
  Current AI safety evaluation and benchmarking frameworks rely on Western-centric culture-agnostic defaults that mask critical regional laws, socio-linguistic nuances, and cultural taboos.<n>We introduce Pluralis v0.1: a novel multimodal, multi-regional, and multilingual dataset built from a culture-first perspective.
  arXiv  Detail & Related papers  (2026-07-07T12:21:52Z)
• Communication Heterogeneity and Collective Consensus in Neural Cellular Automata [0.0]
  We find that linguistic distance consensus produces mild divergence between groups rather than full fragmentation.<n>We admit a natural reading as Ising relaxation, in which a foreign-language region acts as a boundary defect that leaves the system in a higher-energy, partially ordered state.
  arXiv  Detail & Related papers  (2026-06-19T08:15:29Z)
• Multilingual Refusal Alignment for Safer Large Language Models [53.64286756804503]
  We investigate whether single-language alignment transfers cross-lingually, how language consistency is preserved during training, and the resulting trade-offs with general knowledge capabilities.<n>We introduce RefusEU, a novel refusal alignment dataset covering 12 European languages, including a dedicated test set for evaluating current state-of-the-art models.<n>Our controlled Direct Preference Optimization (DPO) experiments provide two key insights: aligning models exclusively in English is insufficient to ensure cross-lingual safety, even for the same harm categories, whereas training on multilingual datasets can improve safety without degrading general performance, as measured by the Global MMLU benchmark.
  arXiv  Detail & Related papers  (2026-04-24T09:29:14Z)
• Bridging the Multilingual Safety Divide: Efficient, Culturally-Aware Alignment for Global South Languages [8.667909336164465]
  Large language models (LLMs) are being deployed across the Global South.<n> Everyday use involves low-resource languages, code-mixing, and culturally specific norms.<n>Our aim is to make multilingual safety a core requirement-not an add-on-for equitable AI in underrepresented regions.
  arXiv  Detail & Related papers  (2026-02-14T19:56:40Z)
• When Meanings Meet: Investigating the Emergence and Quality of Shared Concept Spaces during Multilingual Language Model Training [57.230355403478995]
  We investigate the development of language-agnostic concept spaces during pretraining of EuroLLM.<n>We find that shared concept spaces emerge early and continue to refine, but that alignment with them is language-dependent.<n>In contrast to prior work, our fine-grained manual analysis reveals that some apparent gains in translation quality reflect shifts in behavior.
  arXiv  Detail & Related papers  (2026-01-30T11:23:01Z)
• HarmoCLIP: Harmonizing Global and Regional Representations in Contrastive Vision-Language Models [63.87966115136411]
  HarmoCLIP is a novel framework designed to harmonize global and region representations within Contrastive Language-Image Pre-training.<n>To strengthen the representation capability at the local level, our method introduces a novel Region-Language Alignment supervision strategy.
  arXiv  Detail & Related papers  (2025-11-27T16:24:53Z)
• Multilingual Text-to-Image Person Retrieval via Bidirectional Relation Reasoning and Aligning [81.43257201833154]
  We propose Bi-IRRA: a Bidirectional Implicit Relation Reasoning and Aligning framework to learn alignment across languages and modalities.<n>Within Bi-IRRA, a bidirectional implicit relation reasoning module enables bidirectional prediction of masked image and text.<n>The proposed method achieves new state-of-the-art results on all multilingual TIPR datasets.
  arXiv  Detail & Related papers  (2025-10-20T16:01:11Z)
• MPO: Multilingual Safety Alignment via Reward Gap Optimization [88.76638442683391]
  Large language models (LLMs) have become increasingly central to AI applications worldwide.<n>Existing preference learning methods for safety alignment, such as RLHF and DPO, are primarily monolingual and struggle with noisy multilingual data.<n>We introduce Multilingual reward gaP Optimization (MPO), a novel approach that leverages the well-aligned safety capabilities of the dominant language (English) to improve safety alignment across multiple languages.
  arXiv  Detail & Related papers  (2025-05-22T16:24:51Z)
• MR. Guard: Multilingual Reasoning Guardrail using Curriculum Learning [56.79292318645454]
  Large Language Models (LLMs) are susceptible to adversarial attacks such as jailbreaking. This vulnerability is exacerbated in multilingual setting, where multilingual safety-aligned data are often limited. We propose an approach to build a multilingual guardrail with reasoning.
  arXiv  Detail & Related papers  (2025-04-21T17:15:06Z)
• The Hidden Space of Safety: Understanding Preference-Tuned LLMs in Multilingual context [0.9130277390156759]
  Alignment tuning has enabled large language models to excel in reasoning, instruction-following, and minimizing harmful generations. Despite their widespread deployment, these models exhibit a monolingual bias, raising concerns about the effectiveness of alignment across languages. Current alignment methods predominantly focus on English, leaving it unclear how alignment mechanism generalizes to multilingual settings.
  arXiv  Detail & Related papers  (2025-04-03T15:46:46Z)
• Lens: Rethinking Multilingual Enhancement for Large Language Models [70.85065197789639]
  We propose Lens, a novel approach to enhance multilingual capabilities in large language models (LLMs)<n>Lens operates on two subspaces: the language-agnostic subspace, where it aligns target languages with the central language to inherit strong semantic representations, and the language-specific subspace, where it separates target and central languages to preserve linguistic specificity.<n>Lens significantly improves multilingual performance while maintaining the model's English proficiency, achieving better results with less computational cost compared to existing post-training approaches.
  arXiv  Detail & Related papers  (2024-10-06T08:51:30Z)
• ABC Align: Large Language Model Alignment for Safety & Accuracy [0.0]
  We present ABC Align, a novel alignment methodology for Large Language Models (LLMs) We combine a set of data and methods that build on recent breakthroughs in synthetic data generation, preference optimisation, and post-training model quantisation. Our unified approach mitigates bias and improves accuracy, while preserving reasoning capability, as measured against standard benchmarks.
  arXiv  Detail & Related papers  (2024-08-01T06:06:25Z)
• Byzantine-Robust Gossip: Insights from a Dual Approach [15.69624587054777]
  This paper investigates Byzantine-resilient algorithms in a decentralized setting, where devices communicate directly with one another. We provide both global and local clipping rules in the special case of average consensus, with tight convergence guarantees.
  arXiv  Detail & Related papers  (2024-05-06T13:22:54Z)
• Unintended Impacts of LLM Alignment on Global Representation [62.6579934112071]
  We show that developers align Large Language Models (LLMs) to user preferences through a variety of procedures, such as Reinforcement Learning From Human Feedback (RLHF) and Direct Preference Optimization (DPO) We explore how alignment impacts performance along three axes of global representation: English dialects, multilingualism, and opinions from and about countries worldwide. We conclude by discussing design decisions that led to these unintended impacts and recommendations for more equitable preference tuning.
  arXiv  Detail & Related papers  (2024-02-22T23:31:22Z)
• Adaptive Global-Local Representation Learning and Selection for Cross-Domain Facial Expression Recognition [54.334773598942775]
  Domain shift poses a significant challenge in Cross-Domain Facial Expression Recognition (CD-FER) We propose an Adaptive Global-Local Representation Learning and Selection framework.
  arXiv  Detail & Related papers  (2024-01-20T02:21:41Z)
• Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning [60.058083574671834]
  This paper presents a novel FCCL+, federated correlation and similarity learning with non-target distillation. For heterogeneous issue, we leverage irrelevant unlabeled public data for communication. For catastrophic forgetting in local updating stage, FCCL+ introduces Federated Non Target Distillation.
  arXiv  Detail & Related papers  (2023-09-28T09:32:27Z)
• Evaluating the Diversity, Equity and Inclusion of NLP Technology: A Case Study for Indian Languages [35.86100962711644]
  In order for NLP technology to be widely applicable, fair, and useful, it needs to serve a diverse set of speakers across the world's languages. We propose an evaluation paradigm that assesses NLP technologies across all three dimensions.
  arXiv  Detail & Related papers  (2022-05-25T11:38:04Z)
• GL-CLeF: A Global-Local Contrastive Learning Framework for Cross-lingual Spoken Language Understanding [74.39024160277809]
  We present Global--Local Contrastive Learning Framework (GL-CLeF) to address this shortcoming. Specifically, we employ contrastive learning, leveraging bilingual dictionaries to construct multilingual views of the same utterance. GL-CLeF achieves the best performance and successfully pulls representations of similar sentences across languages closer.
  arXiv  Detail & Related papers  (2022-04-18T13:56:58Z)
• Towards Understanding and Mitigating Social Biases in Language Models [107.82654101403264]
  Large-scale pretrained language models (LMs) can be potentially dangerous in manifesting undesirable representational biases. We propose steps towards mitigating social biases during text generation. Our empirical results and human evaluation demonstrate effectiveness in mitigating bias while retaining crucial contextual information.
  arXiv  Detail & Related papers  (2021-06-24T17:52:43Z)
• Seeking the Shape of Sound: An Adaptive Framework for Learning Voice-Face Association [94.7030305679589]
  We propose a novel framework to jointly address the above-mentioned issues. We introduce a global loss into the modality alignment process. The proposed method outperforms the previous methods in multiple settings.
  arXiv  Detail & Related papers  (2021-03-12T14:10:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.