Search papers, labs, and topics across Lattice.
4
0
4
4
LLMs falter in corporate Q&A tasks, with performance dropping sharply as document complexity increases.
Summaries of literary works often distort the original narrative, with human and LLM annotators both failing to maintain linearity and uniformity in their representations.
LLMs are churning out eerily similar "lighthouse" stories, revealing that alignment data can have a surprisingly outsized impact on generation diversity, even more so than pre-training data.
LLMs focus on the ends of novels when summarizing, unlike humans, revealing a potential bottleneck in long-context narrative understanding.
