Search papers, labs, and topics across Lattice.
2
0
6
2
LLMs often fail to maintain accurate beliefs in multi-turn interactions, but targeted reinforcement learning and representation steering can dramatically improve their contextual reasoning.
Stealing just the right neurons from another LLM lets you patch safety holes or remove biases in your own, with almost no performance hit.
