Search papers, labs, and topics across Lattice.
Affiliation:
2
0
4
Repurposing speculative decoding backbones for output-length prediction slashes short-request latency by nearly 35%.
LLMs struggle with complex temporal reasoning, showing significant accuracy drops on intricate waveform queries, underscoring a critical gap in their capabilities.
