← back to paper

Review history

arxiv: 2606.08761 · 2 revisions

APEX4: Efficient Pure W4A4 LLM Inference via Intra-SM Compute Rebalancing

1. 2026-06-30 CONDITIONAL LOW v0.9.1-grok novelty 7.0
   38577 ms 5895 in 1365 out 2026-06-30T11:05:48.517258+00:00
2. 2026-06-27 UNVERDICTED LOW v0.9.1-grok novelty 6.0
   16459 ms 5871 in 1409 out 2026-06-27T17:46:26.206576+00:00