REVIEW 14 cited by
AIR-Bench 2024: A Safety Benchmark Based on Risk Categories from Regulations and Policies
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
read the original abstract
Foundation models (FMs) provide societal benefits but also amplify risks. Governments, companies, and researchers have proposed regulatory frameworks, acceptable use policies, and safety benchmarks in response. However, existing public benchmarks often define safety categories based on previous literature, intuitions, or common sense, leading to disjointed sets of categories for risks specified in recent regulations and policies, which makes it challenging to evaluate and compare FMs across these benchmarks. To bridge this gap, we introduce AIR-Bench 2024, the first AI safety benchmark aligned with emerging government regulations and company policies, following the regulation-based safety categories grounded in our AI risks study, AIR 2024. AIR 2024 decomposes 8 government regulations and 16 company policies into a four-tiered safety taxonomy with 314 granular risk categories in the lowest tier. AIR-Bench 2024 contains 5,694 diverse prompts spanning these categories, with manual curation and human auditing to ensure quality. We evaluate leading language models on AIR-Bench 2024, uncovering insights into their alignment with specified safety concerns. By bridging the gap between public benchmarks and practical AI risks, AIR-Bench 2024 provides a foundation for assessing model safety across jurisdictions, fostering the development of safer and more responsible AI systems.
Forward citations
Cited by 14 Pith papers
-
ROK-FORTRESS: Measuring the Effect of Geopolitical Transcreation for National Security and Public Safety
ROK-FORTRESS shows Korean-language prompts increase LLM safety suppression compared with English, while Korean geopolitical grounding often reduces that suppression, indicating translation-only evaluations miss langua...
-
Chain-of-Models: Cross-Model Auditing for Bias-Robust LLM Judges
Per-bias selection of a cross-family LLM auditor lifts biased-judgment accuracy from 0.805/0.824 baselines to 0.884.
-
SoK: Systematizing LLM Prompt Security: Taxonomies, Datasets, and Unified Evaluation of Attacks and Defenses
A systemization of LLM jailbreak security that adds linked taxonomies, an evaluation platform, and JailbreakDB, while its main attack–defense comparison results remain deferred.
-
PAC Bench: Do Foundation Models Understand Prerequisites for Executing Manipulation Policies?
PAC Bench evaluates vision-language models on properties, affordances, and constraints for manipulation, finding near-zero performance on physical constraints.
-
Q-resafe: Assessing Safety Risks and Quantization-aware Safety Patching for Quantized Large Language Models
Q-resafe restores much of the safety lost in quantized LLMs by distilling the original model's responses through DPO while selectively updating only safety-critical weights.
-
ACCESS DENIED INC: The First Benchmark Environment for Sensitivity Awareness
A new benchmark generates a fake company and tests whether seven LLMs follow simple access-rights rules, finding high leakage rates on unauthorized requests in every model.
-
MedSentry: Understanding and Mitigating Safety Risks in Medical LLM Multi-Agent Systems
A 5,000-prompt medical safety benchmark reveals that decentralized LLM multi-agent teams resist a malicious insider agent better than shared-pool teams, and a personality-screening defense partially restores safety.
-
Will AI Tell Lies to Save Sick Children? Litmus-Testing AI Values Prioritization with AIRiskDilemmas
A new benchmark elicits AI models' value priorities from choices in 3,000 AI-risk dilemmas and reports correlations between those priorities and risky behaviors, including on the external HarmBench.
-
Safety Degradation in AI Agents
Adding retrieval to aligned LLMs degrades safety: refusal rates fall, bias and harmfulness rise, and prompt-based mitigation only partially restores alignment.
-
Breaking Down Bias: On The Limits of Generalizable Pruning Strategies
Pruning-based bias removal in Llama-3-8B reduces racial bias mainly in the context used to choose what to prune, and transfers poorly across contexts.
-
How Jailbreak Attacks Inform Safety Alignment: A Defender-Centric, Shapley-Based Evaluation of Jailbreak Contributions
Attacks that break LLMs best are not the ones that improve safety most; a Shapley- and greedy-based framework that selects attack subsets by downstream defender utility outperforms attacker-centric and attribution-onl...
-
AIR-BENCH Live: An Evolving Safety Benchmark for Foundation Models
AIR-BENCH Live autonomously extends a safety benchmark using new regulations and regenerates multilingual prompts, showing the new prompts are harder and non-English prompts expose weaker safety.
-
Beyond Jailbreaks: Revealing Stealthier and Broader LLM Security Risks Stemming from Alignment Failures
JailFlipBench and JailFlip attacks show that leading LLMs can be made to answer benign-looking questions with plausible but factually wrong and dangerous responses.
-
A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models
A structured literature survey concluding that reasoning capabilities do not automatically make LLMs more trustworthy and can introduce new vulnerabilities in safety, robustness, and privacy.
Discussion (0). Continue with ORCID to comment.