Search papers, labs, and topics across Lattice.
IIIT Delhi, India
2
0
5
0
Scaling up LLMs doesn't uniformly improve context handling; instead, it paradoxically amplifies the tendency to copy irrelevant tokens while simultaneously improving resistance to misinformation.
Existing citation recommendation benchmarks overestimate real-world performance because they fail to account for the temporal constraints of recommending citations for *new* papers.
