←back to thread

Over fifty new hallucinations in ICLR 2026 submissions

(gptzero.me)
504 points puttycat | 1 comments | 07 Dec 25 13:16 UTC | HN request time: 3.065s | source
Show context
TaupeRanger ◴[07 Dec 25 14:51 UTC] No.46182062[source]▶
It's going to be even worse than 50:

> Given that we've only scanned 300 out of 20,000 submissions, we estimate that we will find 100s of hallucinated papers in the coming days.

replies(1): >>46182080 #
shusaku ◴[07 Dec 25 14:53 UTC] No.46182080[source]▶
20,000 submissions to a single conference? That is nuts
replies(3): >>46182151 #>>46182203 #>>46182579 #
zipy124 ◴[07 Dec 25 15:51 UTC] No.46182579[source]▶
When academics are graded based on number of papers this is the result.
replies(1): >>46183230 #
adestefan ◴[07 Dec 25 17:11 UTC] No.46183230[source]▶
The problem isn't only papers it's that the world of academic computer science coalesced around conference submissions instead of journal submissions. This isn't new and was an issue 30 years ago when I was in grad school. It makes the work of conference organizes the little block holding up the entire system.
replies(1): >>46183839 #
1. DonaldPShimoda ◴[07 Dec 25 18:28 UTC] No.46183839[source]▶
Makes me grateful I'm in an area of CS where the "big" conferences are like 500 attendees.