6 ms·

Interestingly ChatGPT right now answered > Bixonimania is not a real disease. It was deliberately invented by scientists as an experiment to test whether AI sy

• latexr 5mo ago
  It’s not that interesting, we know companies react to these things fast. It’s why I don’t share online my methods on how simple it is to show LLM flaws. The problem is all the lies which won’t be fessed up to. This one was because they had to to prove the point, but the bad actors with ulterior motives won’t reveal what they’re doing.
  • rcxdude 5mo ago
    Doesn't even need the companies to react fast. Now that the Google results are returning news articles on it the LLMs are going to find and report on that as opposed to the original paper.
• rcxdude 5mo ago
  The news articles on it are going to affect this. I wonder if the original paper is in the base models at all, almost certainly these results were from the article showing up in an Internet search. Similarly, I wonder what a frontier model would say if just given the paper in isolation and asked to summarise/opine on it. I suspect they would successfully recognize such obivous signs, the failure is when less sophisticated LLMs are just skimming search results and summarising them.