Research map / Large language models / Language, safety and interpretability
ExpGuard: LLM Content Moderation in Specialized Domains
Minseok Choi, Dongjin Kim, Seungbin Yang, Subin Kim, Youngjun Kwak, Juyoung Oh, Jaegul Choo, Jungmin Son
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add domain-specific data to the training set (ablation) | ExpGuard · ExpGuardTest Total | prompt F1 (%): 85.3 -> 93.3 | Improved |
| add in-the-wild data to the training set (ablation) | ExpGuard · ExpGuardTest Total | prompt F1 (%): 93.2 -> 93.3 | No clear effect |
| add in-the-wild data to the training set (ablation) | ExpGuard · Public benchmarks (prompt, average) | prompt F1 (%): 84.1 -> 85.7 | Improved |
| add human-written data to the training set (ablation) | ExpGuard · Public benchmarks (response, average) | response F1 (%): 73.9 -> 78.5 | Improved |
| train ExpGuard with domain-specific data versus WildGuard | ExpGuard vs WildGuard · ExpGuardTest Total | prompt F1 (%): 84.4 -> 93.3 | Improved |
| train ExpGuard versus WildGuard | ExpGuard vs WildGuard · ExpGuardTest Total | response F1 (%): 77.4 -> 92.7 | Improved |
| ExpGuard versus WildGuard on public benchmarks | ExpGuard · Public benchmarks (prompt, average) | prompt F1 (%): 84.2 -> 85.7 | Improved |
| ExpGuard versus WildGuard on public benchmark responses | ExpGuard · Public benchmarks (response, average) | response F1 (%): 78.8 -> 78.5 | No clear effect |
| add 270 domain-specific jailbreak prompts to training (ExpGuard+) (ablation) | ExpGuard · Domain-specific jailbreaks | detection rate | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- WildGuard — WildGuard: Open One-stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs (NeurIPS 2024) · better on 9 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PolyGuard: Massive Multi-Domain Safety Policy-Grounded Guardrail Dataset (NeurIPS 2025 · similarity 0.90)
- On Calibration of LLM-based Guard Models for Reliable Content Moderation (ICLR 2025 · similarity 0.90)
- CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs (NeurIPS 2025 · similarity 0.90)
- RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards (NeurIPS 2025 · similarity 0.90)
- WildGuard: Open One-stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs (NeurIPS 2024 · similarity 0.90)
- RigorLLM: Resilient Guardrails for Large Language Models against Undesired Content (ICML 2024 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
