BELLS-O is the first vendor-neutral operational benchmark comparing specialized guardrails and repurposed frontier LLMs on accuracy, false-positive rates, speed, and monetary cost across 11 harm categories and 13 jailbreak techniques.
Granite guardian,
12 Pith papers cite this work. Polarity classification is still indexing.
12
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 12roles
background 1polarities
background 1representative citing papers
RAGognizer adds a detection head to LLMs for joint training on generation and token-level hallucination detection, yielding SOTA detection and fewer hallucinations in RAG while preserving output quality.
Benign fine-tuning collapses safety geometry in guard models like Granite Guardian, dropping refusal to 0%, but Fisher-Weighted Safety Subspace Regularization restores it to 75% while improving robustness.
Disentangled Safety Adapters decouple safety computations from task-optimized LLMs via lightweight adapters, yielding up to 53% better AUC on safety tasks and dynamic inference-time alignment with reduced performance trade-offs.
Cognitive Firewall applies four gates (intent, zero-trust context, consistency, output risk) via an oversight model to cut jailbreak success to 2% or below on most tested sets while keeping over-refusal at 8%.
SkillGuard-Robust formulates pre-load auditing of untrusted Agent Skills as a three-way classification task and achieves 97.30% exact match and 98.33% malicious-risk recall on held-out benchmarks.
Guardian-as-an-Advisor prepends risk labels and explanations from a guardian model to queries, improving LLM safety compliance and reducing over-refusal while adding minimal compute overhead.
Bielik Guard delivers compact Polish safety classifiers with F1 scores near 0.79 and superior real-prompt precision over baselines.
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.
Fine-tuned LLMs produce incoherent safety responses and yield benchmark-dependent conclusions unless evaluations are grounded in explicit capability targets.
Opir introduces efficient multi-task encoder models trained on a 996-category safety taxonomy that match or exceed larger baselines on most safety benchmarks while using under 100M parameters for edge variants.
TWGuard achieves +0.289 F1 improvement and 94.9% false-positive reduction for LLM safety guardrails in the Taiwan linguistic context compared to foundation models and baselines.
citing papers explorer
-
BELLS-O: Evaluating the Operational Trade-offs of LLM Supervision Systems
BELLS-O is the first vendor-neutral operational benchmark comparing specialized guardrails and repurposed frontier LLMs on accuracy, false-positive rates, speed, and monetary cost across 11 harm categories and 13 jailbreak techniques.
-
RAGognizer: Hallucination-Aware Fine-Tuning via Detection Head Integration
RAGognizer adds a detection head to LLMs for joint training on generation and token-level hallucination detection, yielding SOTA detection and fewer hallucinations in RAG while preserving output quality.
-
When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models
Benign fine-tuning collapses safety geometry in guard models like Granite Guardian, dropping refusal to 0%, but Fisher-Weighted Safety Subspace Regularization restores it to 75% while improving robustness.
-
Disentangled Safety Adapters Enable Efficient Guardrails and Flexible Inference-Time Alignment
Disentangled Safety Adapters decouple safety computations from task-optimized LLMs via lightweight adapters, yielding up to 53% better AUC on safety tasks and dynamic inference-time alignment with reduced performance trade-offs.
-
Cognitive Firewall: A Proactive, Zero-Trust, Multi-Gate Framework for LLM Safety
Cognitive Firewall applies four gates (intent, zero-trust context, consistency, output risk) via an oversight model to cut jailbreak success to 2% or below on most tested sets while keeping over-refusal at 8%.
-
Structured Security Auditing and Robustness Enhancement for Untrusted Agent Skills
SkillGuard-Robust formulates pre-load auditing of untrusted Agent Skills as a three-way classification task and achieves 97.30% exact match and 98.33% malicious-risk recall on held-out benchmarks.
-
Guardian-as-an-Advisor: Advancing Next-Generation Guardian Models for Trustworthy LLMs
Guardian-as-an-Advisor prepends risk labels and explanations from a guardian model to queries, improving LLM safety compliance and reducing over-refusal while adding minimal compute overhead.
-
Bielik Guard: Efficient Polish Language Safety Classifiers for LLM Content Moderation
Bielik Guard delivers compact Polish safety classifiers with F1 scores near 0.79 and superior real-prompt precision over baselines.
-
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.
-
Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability
Fine-tuned LLMs produce incoherent safety responses and yield benchmark-dependent conclusions unless evaluations are grounded in explicit capability targets.
-
Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content
Opir introduces efficient multi-task encoder models trained on a 996-category safety taxonomy that match or exceed larger baselines on most safety benchmarks while using under 100M parameters for edge variants.
-
TWGuard: A Case Study of LLM Safety Guardrails for Localized Linguistic Contexts
TWGuard achieves +0.289 F1 improvement and 94.9% false-positive reduction for LLM safety guardrails in the Taiwan linguistic context compared to foundation models and baselines.