Skip to main content
Search

# AutoIndexer

1 related articles

Subscribe to updates
  1. arXiv

    A causal model learns to edit its own draft: insert, replace, delete without recomputing, 59.2 on HumanEval

    Autoregressive decoding commits to its output in generation order and cannot revise it, while diffusion models can correct tokens but stay bound to fixed length and absolute positions. AutoIndexer decouples generation order from display order, using a marker head to decide when to edit and a pointer head to localize the edit span, with perturbation training that teaches the model to undo insertions, substitutions, and deletions. On a Qwen3 8B backbone it preserves standard benchmark performance and scores 59.2 on HumanEval without instructions, above the matched LoRA baseline's 51.2 and the diffusion baselines.

You’re all caught up