evilscript

7 ms·

1. 1.▲
   by evilscript 21d ago

   Seems like there is a promising future for models that can fit on mobile phones (if correctly quantized, I guess). Big up for a small European lab!
2. 2.▲
   How sure is the activation oracle?(arxiv.org)

   1 points by evilscript 4mo ago|0 comments
3. 3.▲
   The Moltbook Files: A Harmless Slopocalypse or Humanity's Last Experiment(arxiv.org)

   4 points by evilscript 4mo ago|0 comments
4. 4.▲
   Instructions in papers can manipulate AI reviewers 78-86% of the time(researchsquare.com)

   3 points by evilscript 8mo ago|3 comments
5. 5.▲
   by evilscript 8mo ago

   We should probably be more worried about researchers gaming AI peer review. We spent weeks running 5,600 experiments where we hid (novel, skeleton-key like) prompt injections inside academic papers. Then we fed them to ChatGPT and Gemini ac
6. 6.▲
   by evilscript 1y ago

   I relate to this article but I find that for a moment it confused verbosity with narrative depth. A piece of text can be very small, almost “compressed”, but doesn’t lose narrative power. This is why I feel like we are not losing verbosity,