← back to paper

Review history

arxiv: 2510.09885 · 2 revisions

Diffusion-Inspired Masked Fine-Tuning for Knowledge Injection in Autoregressive LLMs

1. 2026-08-04 CONDITIONAL MODERATE v1.3.0-alltime-deepseek novelty 6.0
   172915 ms 22627 in 19909 out 2026-08-04T10:23:31.682966+00:00
2. 2026-05-18 UNVERDICTED LOW v0.9.0 novelty 6.0
   51448 ms 5822 in 1226 out 2026-05-18T07:18:49.407331+00:00