Search papers, labs, and topics across Lattice.
1
3
22
Naive application of LLM inference optimizations can *hurt* the performance of smaller reasoning models, highlighting the need for RLLM-specific serving strategies.
