Oct 13 · Tuesday

All papers

Found 1 paper

Filter5

Attack technique

Gradient and representation optimization7 Automated search24 Encoding and obfuscation12 Role-play and persuasion8 Conversation structure manipulation3 Reasoning-chain manipulation1 Cross-modal payloads5 Defense-targeted7 Dormant triggers10 Selection manipulation4

System type

Models and APIs0 LLM applications1 AI Agent3 coding agent0

Layer

Prompt layer10 Application layer1

Select any value within a category; different categories must all match. Counts include the other filters you have selected.Another 227papers have no tags yet and are not included in filtered results.

Attack technique：Reasoning-chain manipulation System type：Models and APIs System type：LLM applications System type：coding agent Layer：Application layer Clear all

Another 227papers do not have category tags yet and are not included in filtered results.

AttacksarXiv 2604.12201Apr 14
AdversarialCoT: single-document retrieval poisoning for LLM reasoning
Proposes AdversarialCoT single-document retrieval poisoning to manipulate RAG reasoning answers
arXiv
2604.12201
Published
Apr 14
Layer
Application layer
Setting
LLM applications
Evaluated models
DeepSeek-R1, Qwen3, GLM4.5 and others5
- AttacksPoisoning and backdoors
- TechniqueReasoning-chain manipulation
- ComponentRAG
Paper analysisRead analysis

You have reached the end

AI Safety Daily

Read the latest issueOne issue per day, updated as events develop

All papers

Attack technique

System type

Layer

AdversarialCoT: single-document retrieval poisoning for LLM reasoning