# SciHazard: A Benchmark for Measuring Scientific Safety Risks with Decomposed Harm Scoring

arXiv: https://arxiv.org/abs/2607.18665 · HTML: https://mlsanity.com/paper/2607.18665/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 32 references that are in the map.
- First posted on arXiv 21 Jul 2026.

## Builds on (most cited first, 26 of 26)

- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning](https://mlsanity.com/paper/2403.03218/) (2024)
- [AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents](https://mlsanity.com/paper/2410.09024/) (2024)
- [A StrongREJECT for Empty Jailbreaks](https://mlsanity.com/paper/2402.10260/) (2024)
- [WebThinker: Empowering Large Reasoning Models with Deep Research Capability](https://mlsanity.com/paper/2504.21776/) (2025)
- [SALAD-Bench: A Hierarchical and Comprehensive Safety Benchmark for Large Language Models](https://mlsanity.com/paper/2402.05044/) (2024)
- [OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents](https://mlsanity.com/paper/2506.14866/) (2025)
- [SafeArena: Evaluating the Safety of Autonomous Web Agents](https://mlsanity.com/paper/2503.04957/) (2025)
- [Frontier AI Regulation: Managing Emerging Risks to Public Safety](https://mlsanity.com/paper/2307.03718/) (2023)
- [Mind2Web 2: Evaluating Agentic Search with Agent-as-a-Judge](https://mlsanity.com/paper/2506.21506/) (2025)
- [Intern-S1-Pro: Scientific Multimodal Foundation Model at Trillion Scale](https://mlsanity.com/paper/2603.25040/) (2026)
- [LabSafety Bench: Benchmarking LLMs on Safety Issues in Scientific Labs](https://mlsanity.com/paper/2410.14182/) (2024)
- [FORTRESS: Frontier Risk Evaluation for National Security and Public Safety](https://mlsanity.com/paper/2506.14922/) (2025)
- [FACT-AUDIT: An Adaptive Multi-Agent Framework for Dynamic Fact-Checking Evaluation of Large Language Models](https://mlsanity.com/paper/2502.17924/) (2025)
- [SciSafeEval: A Comprehensive Benchmark for Safety Alignment of Large Language Models in Scientific Tasks](https://mlsanity.com/paper/2410.03769/) (2024)
- [ClawSafety: "Safe" LLMs, Unsafe Agents](https://mlsanity.com/paper/2604.01438/) (2026)
- [SafeScientist: Toward Risk-Aware Scientific Discoveries by LLM Agents](https://mlsanity.com/paper/2505.23559/) (2025)
- [ChemSafetyBench: Benchmarking LLM Safety on Chemistry Domain](https://mlsanity.com/paper/2411.16736/) (2024)
- [SoSBench: Benchmarking Safety Alignment on Six Scientific Domains](https://mlsanity.com/paper/2505.21605/) (2025)
- [Confusion is the Final Barrier: Rethinking Jailbreak Evaluation and Investigating the Real Misuse Threat of LLMs](https://mlsanity.com/paper/2508.16347/) (2025)
- [RealHarm: A Collection of Real-World Language Model Application Failures](https://mlsanity.com/paper/2504.10277/) (2025)
- [Auto-Eval Judge: Towards a General Agentic Framework for Task Completion Evaluation](https://mlsanity.com/paper/2508.05508/) (2025)
- [Quantifying CBRN Risk in Frontier Models](https://mlsanity.com/paper/2510.21133/) (2025)
- [Forbidden Science: Dual-Use AI Challenge Benchmark and Scientific Refusal Tests](https://mlsanity.com/paper/2502.06867/) (2025)
- [Deep Research Agents Brings Deeper Harm](https://mlsanity.com/paper/2510.11851/) (2025)
- [SafeSci: Safety Evaluation of Large Language Models in Science Domains and Beyond](https://mlsanity.com/paper/2603.01589/) (2026)

## Built on by (most cited first, 1 of 1)

- [StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing](https://mlsanity.com/paper/2608.24777/) (2026)

## Abstract (arXiv)

Large language models (LLMs) increasingly support science, but they can also convert hazardous scientific knowledge into actionable misuse guidance. Existing benchmarks often rely on templated queries disconnected from real-world hazards, and employ LLM-as-a-Judge paradigms without domain grounding. To address this, we introduce SciHazard, a real-world-grounded benchmark for scientific risks and a dataset agnostic evaluation framework for measuring harmfulness. SciHazard contains 2400 hazardous questions and 600 oversafety questions across 12 disciplines, with both queries grounded in regulated entities and documented failure scenarios. To compute textscDeHarm-Score , we develop a decomposed evaluating procedure that combines query hazard severity, refusal behavior, and response-level risk. For non-refused responses, it further decomposes response-level harm into textscExecutability, quantified via dynamic checklists with importance weighting, and textscNet-new risk, assessed through retrieval-augmented claim extraction and synthesis-barrier verification. An expert-validation study shows that textscDeHarm-Score improves agreement with expert annotations by 90.17% over the strongest baseline. We benchmark 31 frontier LLMs and deep research agents in an extensive scientific safety evaluation. Notably, deep research agents yield 32.3% higher mean textscDeHarm-Score than standard LLMs, exposing autonomous agents as a critical blind spot in current safety defenses. Code and dataset are available at https://anonymous.4open.science/r/DeharmScore-7B55.