CREST-Search is a red-teaming framework that crafts seemingly benign search queries to induce unsafe citations from web-augmented LLMs, backed by a new WebSearch-Harm dataset for fine-tuning a specialized attacker model.
Llama guard 3-1b-int4: Compact and 9 efficient safeguard for human-ai conversations
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
citation-role summary
other 1
citation-polarity summary
roles
other 1polarities
unclear 1representative citing papers
RecurGuard monitors recurrence rate, volume growth, and query progress in exposed reasoning traces to terminate generation on token-consumption attacks, reporting 99% detection on OverThink and 92% on ExtendAttack with near-zero false positives.
LPG compresses policy deliberation into 10 latent tokens to reach 84.5% safety accuracy and 11x speedup over explicit reasoning baselines on guardrail benchmarks.
LLMs propagate misinformation more in lower-resource languages and lower-HDI countries, with input safety classifiers and retrieval-augmented fact-checking showing cross-lingual and regional gaps.
MobileLLM-Flash creates 350M-1.4B parameter LLMs via latency-guided search and attention skipping, delivering up to 1.8x faster prefill and 1.6x faster decode on mobile CPUs with comparable or better quality.
EvoSynth evolves code-based jailbreak algorithms via multi-agent self-correction, reaching 85.5% ASR on Claude-Sonnet-4.5 and 95.9% average across targets with greater diversity.
A joint prompt-response verification framework using intent analysts, harm analysts, and a judge improves average F1 to 0.95 and cuts attack success rate to 4.1% across jailbreaks, prompt injection, phishing, cyber abuse, and harmful content.
Soft prompt distillation with total variation and KL divergence transfers safety behaviors from guard models to on-device LLMs and outperforms LoRA adapters, steering vectors, and direct optimization in safety-usefulness trade-offs with minimal inference cost.
A compliance-scored best-of-N orchestration layer for multimodal document generation reports 91% compliance at 5 attempts in 20 seconds and +11 percentage point win rate gains in aggregate operational data for payments dispute defense.
citing papers explorer
-
When Search Goes Wrong: Red-Teaming Web-Augmented Large Language Models
CREST-Search is a red-teaming framework that crafts seemingly benign search queries to induce unsafe citations from web-augmented LLMs, backed by a new WebSearch-Harm dataset for fine-tuning a specialized attacker model.
-
RecurGuard: Runtime Monitoring for Reasoning-Token Consumption Attacks
RecurGuard monitors recurrence rate, volume growth, and query progress in exposed reasoning traces to terminate generation on token-consumption attacks, reporting 99% detection on OverThink and 92% on ExtendAttack with near-zero false positives.
-
LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails
LPG compresses policy deliberation into 10 latent tokens to reach 84.5% safety accuracy and 11x speedup over explicit reasoning baselines on guardrail benchmarks.
-
To Lie or Not to Lie? Investigating The Biased Spread of Global Lies by LLMs
LLMs propagate misinformation more in lower-resource languages and lower-HDI countries, with input safety classifiers and retrieval-augmented fact-checking showing cross-lingual and regional gaps.
-
MobileLLM-Flash: Latency-Guided On-Device LLM Design for Industry Scale Deployment
MobileLLM-Flash creates 350M-1.4B parameter LLMs via latency-guided search and attention skipping, delivering up to 1.8x faster prefill and 1.6x faster decode on mobile CPUs with comparable or better quality.
-
Evolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs
EvoSynth evolves code-based jailbreak algorithms via multi-agent self-correction, reaching 85.5% ASR on Claude-Sonnet-4.5 and 95.9% average across targets with greater diversity.
-
Verifying Intent and Harm: A Unified Defense Against LLM-Generated Threats
A joint prompt-response verification framework using intent analysts, harm analysts, and a judge improves average F1 to 0.95 and cuts attack success rate to 4.1% across jailbreaks, prompt injection, phishing, cyber abuse, and harmful content.
-
Distilling Safe LLM Systems via Soft Prompts for On Device Settings
Soft prompt distillation with total variation and KL divergence transfers safety behaviors from guard models to on-device LLMs and outperforms LoRA adapters, steering vectors, and direct optimization in safety-usefulness trade-offs with minimal inference cost.
-
Compliance-Scored Best-of-N Guardrail Orchestration for Multimodal Document Generation in Payments Dispute Defense
A compliance-scored best-of-N orchestration layer for multimodal document generation reports 91% compliance at 5 attempts in 20 seconds and +11 percentage point win rate gains in aggregate operational data for payments dispute defense.