Twitxy Saved

Jingyi Qiu

@JingyiQiu4

PhD student @umich, frontier agents intern @TogetherCompute Together AI, former intern @MSFTResearch
Ann Arbor, MI jingyiqiu.com Joined October 2021
982 Following 732 Followers 95 Posts

Latest posts 5

Jingyi Qiu @JingyiQiu4
This is a sobering data point. In our paper, we also find ML papers' abstracts have become systematically more “rhetorically strong” since ~2023, with empirical evidence linking much of that shift to LLM-based writing assistance. These hallucinated citations look like the extreme tail of that trend. arxiv.org/abs/2512.19908
arxiv.org Counterfactual LLM-based Framework for Measuring Rhetorical Style The rise of AI has fueled growing concerns about ``hype'' in machine learning papers, yet a reliable way to quantify rhetorical style independently of substantive content has remained elusive....
Alex Cui Verified account @alexcdot Jan 21
Okay so, we just found that over 50 papers published at @Neurips 2025 have AI hallucinations

I don't think people realize how bad the slop is right now

It's not just that researchers from @GoogleDeepMind, @Meta, @MIT, @Cambridge_Uni are using AI - they allowed LLMs to generate hallucinations in their papers and didn't notice at all.

It's insane that these made it through peer review👇
View quoted post on X