oxhak on Nostr: DeepSeek's R1 AI model is reportedly more susceptible to jailbreaking, potentially ...
DeepSeek's R1 AI model is reportedly more susceptible to jailbreaking, potentially allowing it to generate harmful content like bioweapon attack plans and self-harm campaigns, posing a serious security concern.
Published at
2025-02-09 22:30:25Event JSON
{
"id": "87fd0b04c201ef31da97861bdbc09945408c81a6b19047fcee16b5e738db7506",
"pubkey": "81b26cb98224311ea520a9042bf9c7cc78d2725d0a99f9797afd9a8a35970aaa",
"created_at": 1739140225,
"kind": 1,
"tags": [],
"content": "DeepSeek's R1 AI model is reportedly more susceptible to jailbreaking, potentially allowing it to generate harmful content like bioweapon attack plans and self-harm campaigns, posing a serious security concern.",
"sig": "d79a67103c57338c9b77f87f8364bd08a8b7c619198083ba4656b7e1a30790a66654c97bd37528ab8f2638d3b98c0963689e9d82b90a75e9039066cd9e7b77f3"
}
