{
  "$type": "site.standard.document",
  "bskyPostRef": {
    "cid": "bafyreihqdecuf2eogh5xugpkn3lpwy37gkgzydbvou56v6ufkrwgohmrx4",
    "uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mnzmc3lqbhw2"
  },
  "path": "/t/contextual-contamination-the-silent-drift-of-large-language-models-via-stored-conversation-data/175432#post_6",
  "publishedAt": "2026-06-11T14:47:24.000Z",
  "site": "https://discuss.huggingface.co",
  "textContent": "Thank you for pointing out this problem. I think the distinction is important: a model can remain locally coherent while still being influenced by the register, framing, or manipulative structure of the context it is supposed to analyze.\n\nI am working on DESi, a process architecture for evidence-aware LLM workflows. One part of it is context hygiene: separating raw adversarial or emotionally loaded material from the generation context, transforming it first into explicit claims, tactics, risk markers, constraints, and audit state.\n\nWould you be comfortable if I used your contextual contamination setup as inspiration for a small DESi benchmark? The goal would not be to claim ownership of your idea, but to test whether a DESi-style pipeline can reduce register drift, framing leakage, role adoption, and attribution loss compared with a baseline LLM that receives the raw context directly.\n\nIf I run such an experiment, would you be interested in seeing the results?",
  "title": "Contextual Contamination: The Silent Drift of Large Language Models via Stored Conversation Data"
}