home.social

#poisoneddocuments — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #poisoneddocuments, aggregated by home.social.

fetched live
  1. A study by #Anthropic, the UK AI Security Institute, and the Alan Turing Institute found that as few as 250 #maliciousdocuments can #backdoor large language models (#LLMs), regardless of size. This challenges the assumption that attackers need a percentage of #trainingdata, suggesting a fixed number of #poisoneddocuments is sufficient. anthropic.com/research/small-s #tech #media #news

  2. A study by #Anthropic, the UK AI Security Institute, and the Alan Turing Institute found that as few as 250 #maliciousdocuments can #backdoor large language models (#LLMs), regardless of size. This challenges the assumption that attackers need a percentage of #trainingdata, suggesting a fixed number of #poisoneddocuments is sufficient. anthropic.com/research/small-s #tech #media #news