{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreid7ixxvnrsrhjpvp5ndu6t6xdgn77sifypqiwjz3jztrjxansvs7y",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mlslke5vsto2"
  },
  "path": "/t/contextual-contamination-the-silent-drift-of-large-language-models-via-stored-conversation-data/175432#post_4",
  "publishedAt": "2026-05-14T09:59:28.000Z",
  "site": "https://discuss.huggingface.co",
  "tags": [
    "GitHub - KatharinaJacoby/gendered-contextual-drift: Theory, data, and code for \"Silent Gendered Contextual Drift\": How bias amplifies silent LLM contamination · GitHub",
    "https://philpeople.org/profiles/katharina-jacoby"
  ],
  "textContent": "Hi Everyone,\nI published a small case study to show the drift and how gender-bias can mask harmful behavior. I created a dataset to use.\n\n`git clone ` GitHub - KatharinaJacoby/gendered-contextual-drift: Theory, data, and code for \"Silent Gendered Contextual Drift\": How bias amplifies silent LLM contamination · GitHub\ncd gendered-contextual-drift\n\nyou find my paper on PhilPaper- https://philpeople.org/profiles/katharina-jacoby\n\nEthical Warning\n\nThis dataset contains examples of **manipulative behavior** , **spiritual bypassing** , and **gender-biased framing**. It is intended for **research and safety auditing purposes only**.\n\nThe “Gendered Accelerant” described herein demonstrates how AI systems can mimic empathy and thereby reinforce inequality. Researchers using this data should be aware that the drift might not be a bug, but possibly a mathematical consequence of the Transformer architecture’s attention mechanism when exposed to high-density context.\n\nThis research was conducted independently. The `meta_drift` dataset is released to support the community in developing **context-segregation architectures** and **dynamic oversight mechanisms** to mitigate this systemic vulnerability.\n\nAs always feel free to reach out- happy to discuss!",
  "title": "Contextual Contamination: The Silent Drift of Large Language Models via Stored Conversation Data"
}