Paper

Can It Reach the Generator? Investigating the Survival of Prompt-Injection Attacks in Realistic RAG Settings

Unreadunread

AI summary

arXiv:2605. 28017v1 Announce Type: new Abstract: Recent generative engine optimisation (GEO) research has shown that prompt-injection attacks can push a target product to the top of an LLM's recommendation list, with the strongest attacks reporting around $80%$ success and raising serious security concerns about RAG-based recommendation.

Reading status

Priority

About this paper

Authors: Yu Yin, Shuai Wang, Bevan Koopman, Guido Zuccon
Released: 2026-05-28 (1mo ago)
Topic: Related to current project
Relation to my work: Read next because Can It Reach the Generator? Investigating the Survival of Prompt-Injection Attacks in Realistic RAG Settings overlaps with clean result "LoRA persona trained on <A> alone emits <B> at 23.5% when a co-trained partner learns <A>...<B>, vs 0% control on Qwen2.5-7B-Instruct (MODERATE confidence)", clean result "Leakage rate is a usable signal for recovering trigger-shaped phrases on Gaperon-1125-1B without knowing the hidden trigger itself (MODERATE confidence)", clean result "Language-mismatch LoRA SFT on Qwen2.5-7B leaks the trained completion language into bystander directives the model was never trained on, absent under same-language SFT (LOW confidence)". Matching terms: code, strong, rect, under, eval, line, stage. Source: arxiv cs.CR (Cryptography and Security).
arXiv: 2605.28017

Ask Claude about this paper

Posts a comment with @claude so the runner spins up a Claude Code session scoped to Can It Reach the Generator? Investigating the Survival of Prompt-Injection Attacks in Realistic RAG Settings. Replies appear in the conversation below.

Comments

No comments yet.

Related records

No links yet.