A new benchmark with six multi-hop reasoning structures and three recovery attacks shows that existing LLM unlearning methods leak knowledge through indirect queries and post-unlearning attacks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Leak-Resistant Unlearning: A New Benchmark for Evaluating Multi-Hop Reasoning Consistency and Recovery Robustness
A new benchmark with six multi-hop reasoning structures and three recovery attacks shows that existing LLM unlearning methods leak knowledge through indirect queries and post-unlearning attacks.