Search papers, labs, and topics across Lattice.
3
0
7
5
LEEPS achieves a remarkable 3.7% performance boost in reasoning tasks by intelligently balancing prompt exploitation and exploration, redefining efficiency in RLVR for large language models.
GraphRAG, thought to be more robust to poisoning attacks due to its KG abstraction, is surprisingly vulnerable to KEPo, a novel attack that forges knowledge evolution paths to inject toxic events.
Safety-aligned LLMs are so consistently risk-averse that a single, transferable "poison" document can now block up to 96% of queries across different RAG systems, even without access to the target model.
