TRACE detects corpus poisoning in RAG via token influence attribution to find recurrent keywords tied to target answers.
Model-Based Safety and Assessment: 9th International Symposium, IMBSA 2025, Athens, Greece, September 24–26
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Tracing Target Answers in Poisoned Retrieval Corpora via Token Influence Attribution
TRACE detects corpus poisoning in RAG via token influence attribution to find recurrent keywords tied to target answers.