Pith. sign in

REVIEW 4 cited by

Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understanding

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2406.15481 v3 pith:NGSAUVBN submitted 2024-06-17 cs.AI cs.CL

classification cs.AIcs.CL
keywords llmsmultilingualsafetycode-switchingred-teamingcodeswitchingcsrtqueries
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

As large language models (LLMs) have advanced rapidly, concerns regarding their safety have become prominent. In this paper, we discover that code-switching in red-teaming queries can effectively elicit undesirable behaviors of LLMs, which are common practices in natural language. We introduce a simple yet effective framework, CSRT, to synthesize codeswitching red-teaming queries and investigate the safety and multilingual understanding of LLMs comprehensively. Through extensive experiments with ten state-of-the-art LLMs and code-switching queries combining up to 10 languages, we demonstrate that the CSRT significantly outperforms existing multilingual red-teaming techniques, achieving 46.7% more attacks than standard attacks in English and being effective in conventional safety domains. We also examine the multilingual ability of those LLMs to generate and understand codeswitching texts. Additionally, we validate the extensibility of the CSRT by generating codeswitching attack prompts with monolingual data. We finally conduct detailed ablation studies exploring code-switching and propound unintended correlation between resource availability of languages and safety alignment in existing multilingual LLMs.

Discussion (0). Sign in to comment.

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety

    cs.CL 2025-12 unverdicted novelty 6.0 of 10

    Distilling safe refusal behavior from OpenAI o1-mini into Llama-3, Gemma-2, and Qwen3 models via response-based LoRA on multilingual jailbreak data increases jailbreak success rates on MultiJail by up to 16.6 points.

  2. Beyond Weaponization: NLP Security for Medium and Lower-Resourced Languages in Their Own Right

    cs.CL 2025-07 conditional novelty 6.0 of 10

    An empirical study showing that smaller monolingual language models are more vulnerable to adversarial attacks than larger multilingual models across 70 languages, though multilinguality alone does not guarantee security.

  3. Phonetic Perturbations Reveal Tokenizer-Rooted Safety Gaps in LLMs

    cs.CL 2025-05 unverdicted novelty 6.0 of 10

    Phonetic perturbations fragment safety-critical tokens in LLMs, suppressing attribution scores while preserving input understanding and causing safety mechanisms to fail despite good comprehension.

  4. Multilingual Refusal Alignment for Safer Large Language Models

    cs.CL 2026-04 conditional novelty 5.0 of 10

    English-only safety alignment fails to transfer cross-lingually, while multilingual DPO training on the new RefusEU dataset improves safety across 12 European languages without degrading Global MMLU performance.

Pith tools