Search papers, labs, and topics across Lattice.
1
0
3
LLM reading assistants don't need to hallucinate to be harmful; they can subtly steal the user's interpretive labor, even when designed with "epistemic guardrails."
