LLM agents often fail to abstain at the right time in uncertain multi-turn tasks, and the CONVOLVE context engineering method raises timely abstention rates on WebShop from 26.7 to 57.4 without parameter updates.
InProceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, pages 431–445, Singapore
12 Pith papers cite this work, alongside 4 external citations. Polarity classification is still indexing.
12
Pith papers citing it
4
external citations · external index
representative citing papers
On a bilingual NSPS benchmark, Korean language consistently lowers harmful compliance (~10pp TRS), Korean grounding often mitigates that drop, and open- vs closed-source models reverse under direct requests.
UbuntuGuard is the first culturally-grounded policy benchmark for African languages, derived from expert-authored queries, showing that English-centric safety evaluations overestimate real-world multilingual performance.
Culturally-adapted red-teaming prompts raise ASR by a mean of 9.3 pp over direct translations across 16 language-model pairs in four Asian languages, with DT scoring mean cultural depth of 0.17 versus up to 2.51 for CA.
AgentShield uses layered deception traps in LLM agent tool interfaces to detect indirect prompt injection compromises with 90.7-100% success on commercial models, zero false positives, and cross-lingual transfer without retraining.
MSD enables cross-lingual safety transfer in LLMs via self-distillation with Dual-Perspective Safety Weighting, improving safety in low-resource languages without target response data.
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.
Introduces NoisyToolBench benchmark and Ask-when-Needed framework to improve LLM tool-use performance when user instructions are unclear or incomplete.
Translating unsafe inputs to low-resource languages jailbreaks GPT-4 at rates on par with or exceeding state-of-the-art attacks.
A multi-role red-teaming framework with attacker, target, and jury LLMs measures faithfulness in English and Arabic, finding false-premise prompts and length limits change unfaithfulness rates.
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.
citing papers explorer
-
Agentic Abstention: Do Agents Know When to Stop Instead of Act?
LLM agents often fail to abstain at the right time in uncertain multi-turn tasks, and the CONVOLVE context engineering method raises timely abstention rates on WebShop from 26.7 to 57.4 without parameter updates.
-
ROK-FORTRESS: Measuring the Effect of Geopolitical Transcreation for National Security and Public Safety
On a bilingual NSPS benchmark, Korean language consistently lowers harmful compliance (~10pp TRS), Korean grounding often mitigates that drop, and open- vs closed-source models reverse under direct requests.
-
UbuntuGuard: A Culturally-Grounded Policy Benchmark for Equitable AI Safety in African Languages
UbuntuGuard is the first culturally-grounded policy benchmark for African languages, derived from expert-authored queries, showing that English-centric safety evaluations overestimate real-world multilingual performance.
-
Culturally-Adapted Red-Teaming Across East and Southeast Asian Contexts: A Methodological and Comparative Analysis
Culturally-adapted red-teaming prompts raise ASR by a mean of 9.3 pp over direct translations across 16 language-model pairs in four Asian languages, with DT scoring mean cultural depth of 0.17 versus up to 2.51 for CA.
-
AgentShield: Deception-based Compromise Detection for Tool-using LLM Agents
AgentShield uses layered deception traps in LLM agent tool interfaces to detect indirect prompt injection compromises with 90.7-100% success on commercial models, zero false positives, and cross-lingual transfer without retraining.
-
Multilingual Safety Alignment via Self-Distillation
MSD enables cross-lingual safety transfer in LLMs via self-distillation with Dual-Perspective Safety Weighting, improving safety in low-resource languages without target response data.
-
Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety
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.
-
Learning to Ask: When LLM Agents Meet Unclear Instruction
Introduces NoisyToolBench benchmark and Ask-when-Needed framework to improve LLM tool-use performance when user instructions are unclear or incomplete.
-
Low-Resource Languages Jailbreak GPT-4
Translating unsafe inputs to low-resource languages jailbreaks GPT-4 at rates on par with or exceeding state-of-the-art attacks.
-
A Red Teaming Framework for Large Language Models: A Case Study on Faithfulness Evaluation
A multi-role red-teaming framework with attacker, target, and jury LLMs measures faithfulness in English and Arabic, finding false-premise prompts and length limits change unfaithfulness rates.
-
Multilingual Refusal Alignment for Safer Large Language Models
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.
- Why Do Safety Guardrails Degrade Across Languages?