Diagnose Before You Refine: Multi-Dimensional Judge Feedback for Query-Efficient Jailbreak Attacks
A new paper reports a query-efficient jailbreak method with 89.8% HarmBench success in 10 queries.
58
Product
HarmBench
- Mentions · 7d0
- Mentions · 30d2
- All mentions2
- First seenOct 8, 2026
- Last mentioned16d ago
Timeline
2 articles · newest firstSearching for "Harmful Refusal": A Psychometric Audit of an AI Safety Benchmark
A psychometric audit finds HarmBench and HELM Safety do not isolate a single harmful-refusal attribute.
48
Appears with
- HELM Safetyproduct · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.