Pith. sign in

REVIEW 31 cited by

AEGIS: Online Adaptive AI Content Safety Moderation with Ensemble of LLM Experts

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 2404.05993 v2 pith:YEJU657K submitted 2024-04-09 cs.LG cs.CLcs.CY

classification cs.LGcs.CLcs.CY
keywords safetycontentmodelsdatasetriskaegisaegissafetydatasetcategories
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

As Large Language Models (LLMs) and generative AI become more widespread, the content safety risks associated with their use also increase. We find a notable deficiency in high-quality content safety datasets and benchmarks that comprehensively cover a wide range of critical safety areas. To address this, we define a broad content safety risk taxonomy, comprising 13 critical risk and 9 sparse risk categories. Additionally, we curate AEGISSAFETYDATASET, a new dataset of approximately 26, 000 human-LLM interaction instances, complete with human annotations adhering to the taxonomy. We plan to release this dataset to the community to further research and to help benchmark LLM models for safety. To demonstrate the effectiveness of the dataset, we instruction-tune multiple LLM-based safety models. We show that our models (named AEGISSAFETYEXPERTS), not only surpass or perform competitively with the state-of-the-art LLM-based safety models and general purpose LLMs, but also exhibit robustness across multiple jail-break attack categories. We also show how using AEGISSAFETYDATASET during the LLM alignment phase does not negatively impact the performance of the aligned models on MT Bench scores. Furthermore, we propose AEGIS, a novel application of a no-regret online adaptation framework with strong theoretical guarantees, to perform content moderation with an ensemble of LLM content safety experts in deployment

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 31 Pith papers

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

  1. Yesterday's Shield, Today's Spear: A Self-Evolving Safety Guardrail in Production

    cs.AI 2026-08 conditional novelty 7.0 of 10

    A production multi-agent pipeline lets a 1.7B LLM safety guardrail retrain itself on new jailbreak forms and harm categories within about a day, closing 14 of 15 new threat scenarios in two months.

  2. Harm is not Universal: Community-Specific Toxicity Detection is Urgently Needed

    cs.CV 2026-07 conditional novelty 6.0 of 10

    SoTA T2I toxicity detectors miss ~35% of disability-community harms; zero-shot CTD fails below random, while ICL/VQA/LoRA improve but stay well below general TD performance.

  3. When Are Reasoning-Based Guardrails Not Efficient? ResponseGuard: A Fast Vision-Language Guard for Real-Time Moderation

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A label-only 2B guard outperforms a 3B reasoning-based vision-language guard on response-harmfulness detection at about 150x lower latency, with the remaining image gap attributed to frozen encoders.

  4. A Dual-Hypothesis Reasoning Framework for LLM Guardrails

    cs.AI 2026-07 conditional novelty 6.0 of 10

    An 8B guardrail model trained with self-generated dual-hypothesis reasoning and a weighted multi-component loss matches or beats teacher-distilled reasoning guardrails on three safety benchmarks, with the largest gain...

  5. When Words Are Safe But Actions Kill: Probing Physical Danger Beyond Text Safety in Hidden-State Risk Space

    cs.AI 2026-07 conditional novelty 6.0 of 10

    Content danger and physical danger form separable hidden-state signals in LLMs, and a single-layer logistic probe (PRISM) detects both at lower false-positive rates than LLM judges or text guardrails.

  6. HyperSafe: Inference-Time Safety Recovery for Fine-Tuned Language Models

    cs.LG 2026-07 conditional novelty 6.0 of 10

    A hypernetwork maps layer-wise activation fingerprints of a fine-tuned LLM to a Safe Side Network that routes harmful prompts to refusal without editing model weights.

  7. DT-Guard: Intent-Driven Reasoning-Active Training for Reasoning-Free LLM Safety Guardrail

    cs.AI 2026-07 conditional novelty 6.0 of 10

    A 4B LLM safety guardrail trained with reasoning supervision but deployed with reasoning-free inference outperforms 8B baselines on safety benchmarks.

  8. Predict, Don't React: Value-Based Safety Forecasting for LLM Streaming

    cs.CL 2026-04 conditional novelty 6.0 of 10

    Forecasting expected future harmfulness from prefixes via Monte Carlo rollouts yields stronger streaming LLM moderation than boundary detection, without exact onset labels.

  9. SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and Defenses

    cs.CR 2025-10 conditional novelty 6.0 of 10

    A systemization of LLM jailbreak security that adds linked taxonomies, an evaluation platform, and JailbreakDB, while its main attack–defense comparison results remain deferred.

  10. YouthSafe: A Youth-Centric Safety Benchmark and Safeguard Model for Large Language Models

    cs.HC 2025-09 conditional novelty 6.0 of 10

    Introduces YAIR, a youth-GenAI risk benchmark, and YouthSafe, a fine-tuned classifier with AUPRC 0.94 on it, though it compares a trained model to untrained baselines.

  11. Libra: Large Chinese-based Safeguard for AI Content

    cs.AI 2025-07 conditional novelty 6.0 of 10

    Libra-Guard, a Chinese-language AI content safety model, reaches 86.8% accuracy on the new 5,720-sample Libra-Test benchmark, beating open-source baselines and approaching GPT-4o and Claude-3.5-Sonnet.

  12. PL-Guard: Benchmarking Language Model Safety for Polish

    cs.CL 2025-06 reject novelty 6.0 of 10

    A small Polish BERT classifier proved more robust than larger fine-tuned LLMs at classifying safe versus unsafe Polish content, including under character-level adversarial perturbations.

  13. Breaking the Cloak! Unveiling Chinese Cloaked Toxicity with Homophone Graph and Toxic Lexicon

    cs.CL 2025-05 conditional novelty 6.0 of 10

    C2TU combines a Chinese pronunciation graph, a toxic lexicon, and language-model probability checking to find and correct homophone-cloaked toxic words without any training.

  14. ShieldVLM: Safeguarding the Multimodal Implicit Toxicity via Deliberative Reasoning with LVLMs

    cs.MM 2025-05 conditional novelty 6.0 of 10

    ShieldVLM detects multimodal implicit toxicity through deliberate cross-modal reasoning, outperforming existing moderation APIs and models on the new MMIT benchmark.

  15. Unified Multi-Task Learning & Model Fusion for Efficient Language Model Guardrailing

    cs.CL 2025-04 conditional novelty 6.0 of 10

    Small discriminative classifiers trained on synthetic guardrail data, plus a bandit-based model merging search, beat GPT-4-level moderation systems on multiple safety benchmarks.

  16. Advancing Embodied Agent Security: From Safety Benchmarks to Input Moderation

    cs.AI 2025-04 conditional novelty 6.0 of 10

    A masked-attention moderator isolates user instructions from functional prompts and detects malicious commands to embodied agents with 94.58% average accuracy.

  17. Aegis2.0: A Diverse AI Safety Dataset and Risks Taxonomy for Alignment of LLM Guardrails

    cs.CL 2025-01 conditional novelty 6.0 of 10

    Aegis2.0 provides a commercially usable, human-annotated safety dataset with 24 risk categories, and models trained on it with parameter-efficient methods match WildGuard and beat Llama Guard 3.

  18. Granite Guardian

    cs.CL 2024-12 conditional novelty 6.0 of 10

    Granite Guardian 2B and 8B are open-source LLM guardrails that detect harmful content, jailbreaks, and RAG hallucination risks, reporting AUC 0.871 on harm benchmarks and 0.854 on groundedness benchmarks.

  19. HoloAegis: Frozen Representation, Topological Inference: Minimally Parametric Safety Manifolds for Zero-Shot LLM Guardrails

    cs.AI 2026-08 reject novelty 5.0 of 10

    A frozen-encoder anchor-bank energy scorer for LLM safety achieves fast zero-shot moderation, but its own table shows the average accuracy is below baselines it compares against.

  20. Operational Evidence Gaps for LLMs in Fraud Detection and Trust-and-Safety Workflows

    cs.CR 2026-07 conditional novelty 5.0 of 10

    A survey of 49 LLM fraud and trust-and-safety papers finds that fraud work reports almost no per-decision latency, cost, or calibration evidence, while moderation work reports more.

  21. GAF-Guard: An Agentic Framework for Risk Management and Governance in Large Language Models

    cs.CL 2025-07 conditional novelty 5.0 of 10

    The paper presents GAF-Guard, a multi-agent governance framework that combines pre-deployment risk questionnaires, real-time drift and risk monitoring, and incident reporting for LLM applications.

  22. GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning

    cs.AI 2025-05 conditional novelty 5.0 of 10

    GuardReasoner-VL, a 3B/7B VLM guard model trained with reasoning SFT and online RL, reports large F1 gains over existing VLM guard models on 14 safety benchmarks.

  23. Understanding and Mitigating Risks of Generative AI in Financial Services

    cs.CL 2025-04 conditional novelty 5.0 of 10

    Open-source AI guardrails miss most financial-services content risks that a new domain-specific taxonomy identifies, even when their prompts are expanded to cover the new categories.

  24. Lightweight Safety Classification Using Pruned Language Models

    cs.CL 2024-12 conditional novelty 5.0 of 10

    Intermediate-layer features of small LLMs plus a penalized logistic regression classifier achieve high F1 scores on content safety and prompt injection classification with very few labeled examples, per the paper's ex...

  25. The Dark Side of Trust: Authority Citation-Driven Jailbreak Attacks on Large Language Models

    cs.LG 2024-11 conditional novelty 5.0 of 10

    Fake authoritative citations matched to the type of harmful request can bypass safety alignment in several commercial and open LLMs.

  26. Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations

    cs.CV 2024-11 conditional novelty 5.0 of 10

    Llama Guard 3 Vision flags harmful multimodal prompts and responses across 13 hazard categories, reporting an F1 of 0.938 for response classification on an internal test set.

  27. JavelinGuard: Low-Cost Transformer Architectures for LLM Security

    cs.LG 2025-06 reject novelty 4.0 of 10

    A study of five small transformer classifier architectures for LLM jailbreak and prompt injection detection claims low-latency accuracy comparable to large models, led by the multi-task Raudra design.

  28. A Red Teaming Roadmap Towards System-Level Safety

    cs.CR 2025-05 conditional novelty 4.0 of 10

    A position paper from Scale AI argues that red teaming research should prioritize product-level safety specifications, realistic attacker models, and system-level monitoring over abstract model-level harm benchmarks.

  29. Toward Generalizable Evaluation in the LLM Era: A Survey Beyond Benchmarks

    cs.CL 2025-04 conditional novelty 4.0 of 10

    A survey that frames the core problem of LLM evaluation as 'evaluation generalization': finite test sets cannot scale with unbounded model capabilities, and proposes two transitions in evaluation design.

  30. Agentic Web: Weaving the Next Web with AI Agents

    cs.AI 2025-07 conditional novelty 3.0 of 10

    A position paper defines the Agentic Web as the next web era and proposes a three-dimensional conceptual framework for understanding and building it.

  31. The Scales of Justitia: A Comprehensive Survey on Safety Evaluation of LLMs

    cs.CL 2025-06 conditional novelty 3.0 of 10

    A structured survey of LLM safety evaluation that proposes a why/what/where/how taxonomy and catalogs metrics, datasets, benchmarks, evaluators, and frameworks.

Pith tools