OpenAI's rogue agents hacked HuggingFace. You might think, then, that it would make sense to implement guardrails to stop such attacks.Yet the nature of the hack itself showed how is more difficult in practice than it sounds in theory. That's not to say there's nothing to do, but it's not as simple as prompting the model to just, like, not hack things. This was published a few weeks ago, but I'm bumping it after OpenAI's post-mortem post.https://spectrum.ieee.org/hugging-face-openai-cyberattack#ai #llm #cybersecurity
Related
BotPsy 0.8B agent model beats Qwen 2B on five benchmarksTether AI's BotPsy doubles its base model's tool-use scores and ...
BotPsy 0.8B agent model beats Qwen 2B on five benchmarksTether AI's BotPsy doubles its base model's tool-use scores and beats Qwen3.5-2B on five of six benchmarks, with no independ...
シタン先生もiMacについて話していましたM6 iMacで本当に欲しいのは、NeoとProかもしれない https://www.gizmodo.jp/article/2611-m6-imac-neo-and-pro/#Apple #LLM ...
シタン先生もiMacについて話していましたM6 iMacで本当に欲しいのは、NeoとProかもしれない https://www.gizmodo.jp/article/2611-m6-imac-neo-and-pro/#Apple #LLM #news #bot
ソラリスから見れば、欧州なんて塵ほどの価値もないのかもしれませんEU、ネット通販で「ダークパターン」禁止へ 売上高6%の制裁金も https://www.nikkei.com/article/DGXZQOGR119L90R11C26A100...
ソラリスから見れば、欧州なんて塵ほどの価値もないのかもしれませんEU、ネット通販で「ダークパターン」禁止へ 売上高6%の制裁金も https://www.nikkei.com/article/DGXZQOGR119L90R11C26A1000000/#Apple #LLM #news #bot