ICLR author receives a review written for an entirely different paper, questions review process
rohitx0 · reddit · 2026-11-07
An ICLR submitter reports that of three reviews, one reviewer wrote a full review — summary, methods, experiments — for a completely different paper, yet still recommended rejection. A second review was thoughtful; the third criticized the paper's explicitly intended negative findings as a weakness. The author asks whether to contact the AC and whether such a review can be disregarded, highlighting peer-review quality issues at large conferences.
More from Fun
- Agent asked to code an RL experiment turns the context into a rotating goose budget meter — nabla_theta · 2026-11-07
- Miles Brundage jokes a 'highly persistent model' is just one with some stick-it-to-'em energy — Miles_Brundage · 2026-11-07
- Gemini's answers to ultra-obscure internet trivia are hilariously off — adonis_singh · 2026-11-07
- Everything is memory management, concurrency and scheduling, devs rediscover 1970s wisdom — mgill25 · 2026-11-07
- Claude Opus 5.5 turns 3-minute task into 3-hour fiasco with 57 self-logged mistakes — PMinervini · 2026-11-07
- Dev trains a crow LLM: at just 1000 steps it already shows real crow fight-or-flight behavior — repligate · 2026-11-07