Search papers, labs, and topics across Lattice.
4
0
8
0
By cleverly initializing sparse attention with on-chip histograms, AdaSplash-2 achieves comparable or better training speed than FlashAttention-2 at moderate-to-high sparsity, unlocking the potential of $\alpha$-entmax for long-context transformers.
LLMs talk too much: HANDRAISER teaches agents to interrupt each other, slashing communication costs by 32% without sacrificing task performance.
LLMs can now learn to control traffic signals more effectively by debating the best course of action, leading to significant reductions in travel time and congestion.
Current autonomous agent benchmarks miss nearly half of safety violations and over 10% of robustness failures because they only check final outputs, a problem Claw-Eval directly addresses.
