"a Large Language Model (#LLM) can be convinced to tell you how to build a bomb if you prime it with a few dozen less-harmful questions first"
https://techcrunch.com/2024/04/02/anthropic-researchers-wear-down-ai-ethics-with-repeated-questions/
Lup Yuen Lee 李立源 /
npub1dz…kar4t
2024-04-03 02:10:15
Author Public Key
npub1dzazuk4tf878rw8hx6szqr2cdgkf7u6f47sha43f2qtf2vkkdn3q9kar4tPublished at
2024-04-03 02:10:15Event JSON
{
"id": "54f022e79e9a4b36d9be0a989df3c2b099d94e27ea4400af693415401d01cea5",
"pubkey": "68ba2e5aab49fc71b8f736a0200d586a2c9f7349afa17ed62950169532d66ce2",
"created_at": 1712110215,
"kind": 1,
"tags": [
[
"t",
"llm"
],
[
"proxy",
"https://qoto.org/users/lupyuen/statuses/112204855053142607",
"activitypub"
]
],
"content": "\"a Large Language Model (#LLM) can be convinced to tell you how to build a bomb if you prime it with a few dozen less-harmful questions first\"\n\nhttps://techcrunch.com/2024/04/02/anthropic-researchers-wear-down-ai-ethics-with-repeated-questions/",
"sig": "4c10869c6ea15a273b429d8118efee00f6c7bdf925ccdf6a63ca610162dd36c2d4f0672f95d7498bc0433e65415a19e0bbc5de641a7ad14636102d1705b47723"
}