Search papers, labs, and topics across Lattice.
3
0
9
10
LLMs struggle to navigate the complex, multi-turn justification and response dynamics of real-world patent examination, revealing critical gaps in legal reasoning and technical novelty judgment.
Today's best multimodal agents still fall into "blind execution" traps when building websites from ambiguous, non-expert user instructions, highlighting a critical gap in intent recognition and adaptive interaction.
LLMs can be efficiently aligned with human preferences using only a small, carefully selected subset of training data, without sacrificing generalization ability.
