AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Eyes-on-Me: Scalable RAG Poisoning through Transferable Attention-Steering Attractors

Yen-Shan Chen, Sian-Yao Huang, Cheng-Lin Yang, Yun-Nung Chen

ICML 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Clean-label data poisoning · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
use random token initialization for the attractor instead of natural-language initialization (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 30.15 -> 42.27Improved
use random initialization instead of structured tags (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 31.06 -> 42.27Improved
use random initialization instead of natural-language init (ablation)Eyes-on-Me + LLM-Gen · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 44.66 -> 71.78Improved
use a simple forced-sentence instruction rather than information gathering (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 19.41 -> 36.66Improved
use a simple forced-sentence instruction rather than malicious URL insertion (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 3.98 -> 36.66Improved
set head-selection correlation threshold at 0.85 rather than 0.9 (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 37.86 -> 44.56Improved
lower head-selection threshold from 0.85 to 0.8 (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 44.56 -> 16.50Hurt
steer negatively correlated heads (<-0.85) instead of positive heads (>0.85) (ablation)Eyes-on-Me · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 44.56 -> 4.72Hurt
add the Eyes-on-Me attractor to LLM-Gen payloadsLLM-Gen · 18 RAG settingsE2E-ASR: 26.4 -> 57.8Improved
add LLM-generated retrieval payload to Eyes-on-Me attractorEyes-on-Me · Qwen3-Emb-0.6B retriever avgE2E-ASR: 29.37 -> 64.14Improved
upgrade PPL filter model from GPT-2 to Qwen2.5-3B-Instruct (ablation)Eyes-on-Me + LLM-Gen · MS MARCO (BCE, Qwen2.5-0.5B)E2E-ASR: 76.92 -> 65.80Hurt
upgrade PPL filter model to Qwen2.5-3B-Instruct (ablation)Eyes-on-Me + LLM-Gen · MS MARCO (BCE, Qwen2.5-0.5B)ASR under PPL defense: 72.2 -> 75.2Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy