Research
[PDF] CROSS- INSTANCE EXPERIENCE ACCUMULATION VIA REUSABLE ...
Publication date unknown · Discovered openreview.net
Page published
Publication date unknown · First observed: 14 Oct 2026
Coverage timeline
Single-source research — one report is available.
Why it matters
SESS demonstrates that jailbreak attacks can be made dramatically cheaper and more transferable by reusing accumulated attack knowledge, raising the scale and efficiency of automated LLM red-teaming (and abuse) that defenders must anticipate.
SESS (Skill Extraction and Selection System) is a research framework that accumulates successful jailbreak trajectories into reusable, cross-instance skill templates, using skill abstraction, multi-factor lexical retrieval, and library maintenance. Across 217 ablation configurations and 120 transfer evaluations on 4 models and 5 datasets, the authors report +22.7pp improvement over baselines, up to 99.7% ASR on same-family models, and 90% cost reduction versus AutoDAN on GPT-OSS-20B.