Pith. sign in

← back to paper

Review history

arxiv: 2509.08778 · 2 revisions

Do All Autoregressive Transformers Remember Facts the Same Way? A Cross-Architecture Analysis of Recall Mechanisms

  1. 2026-08-15 CONDITIONAL MODERATE v1.4.0-alltime-deepseek-medium novelty 7.0
    86955 ms 17301 in 10968 out 2026-08-15T15:59:02.413359+00:00
  2. 2026-08-04 CONDITIONAL MODERATE v1.3.0-alltime-deepseek novelty 5.0
    89753 ms 17283 in 9756 out 2026-08-04T20:09:51.909978+00:00