{
  "id": "anthropic-sitemap:research:small-samples-poison",
  "type": "article",
  "title": "A small number of samples can poison LLMs of any size",
  "abstract": "_In a joint study with the UK AI Security Institute and the Alan Turing Institute, we found that as few as 250 malicious documents can produce a \"backdoor\" vulnerability in a large language model—regardless of model size or training data volume. Although a 13B parameter model is trained on over 20 times more training data than a 600M model, both can be backdoored by the same small number of poisoned documents. Our results challenge the common assumption that attackers need to control a percentage of training data; instead, they may just need a small, fixed amount. Our study focuses on a narrow backdoor (producing gibberish text) that is unlikely to pose significant risks in frontier models. Nevertheless, we’re sharing these findings to show that data-poisoning attacks might be more practical than believed, and to encourage further research on data poisoning and potential defenses against it._",
  "issued": {
    "date-parts": [
      [
        2025,
        10,
        9
      ]
    ]
  },
  "URL": "https://www.anthropic.com/research/small-samples-poison",
  "publisher": "Anthropic",
  "source": "vendor/anthropic-sitemap/research/small-samples-poison.md"
}