auto-paper-improvement-loop

Iteratively polish a LaTeX paper by running a multi-round GPT review that applies prioritized fixes and recompiles the manuscript to validate changes. This loop refines the document until it reaches submission-ready quality.

Automated LaTeX paper improvement handles reviewer bias by using fresh, context-naive reviewer threads for each round. This independence prevents prior feedback from drifting the critique and maintains regression safety across the document.

Yes, you can restrict edits during automated paper polishing by enforcing an edit whitelist. This optional guardrail prevents unauthorized changes such as adding new citations, altering theorem environments, or modifying the bibliography.

Starting automated LaTeX paper polishing requires a compiled paper directory containing all section .tex files and the compiled PDF. This baseline output from prior plan, figure, write, and compile stages is necessary for the process.

A LaTeX paper improvement loop fails to catch restatement regressions when it skips restatement regression checks during the recompile phase. Running these checks after applying fixes is required to reduce drift and verify document consistency.