Search papers, labs, and topics across Lattice.
1
0
2
LLM-based recommendation gets a boost: initializing item embeddings with semantic keywords and aligning tokens to item clusters significantly improves performance.
