ICLR 2027 AC desk-rejects 5 of 14 submissions over basic writing flaws, no LLM detectors used

xuanalogue · x · 2026-10-19

An ICLR 2027 Area Chair shares why he recommended desk rejections for 5 of 14 submissions — without using any LLM-detection tools, which he argues are unnecessary and shouldn't be used for desk-reject decisions. The papers shared shockingly basic problems, each identifiable in about 10 minutes: undefined jargon (papers must be self-contained), fake or missing result analysis (repeating numbers isn't insight), and missing citations (fewer than three citations in an entire introduction). A word cloud of recurring issues includes poor writing, incoherent logic, weak motivation, and unclear research gaps.

Original post →

More from Fun

Fun channel →