[untitled]2 ポイント·投稿者 coonr·5 か月前·0 コメント4 コメントコメントを投稿[–]ddmma·5 か月前返信Good hygiene. But the gap they admitted—NL attacks, prompt injection—won't get fixed by pointing another LLM at it.LLM auditing LLM is like RAG solving reasoning. Same blindspot, twice.[–]Jang-woo·5 か月前返信It’s encouraging to see the conversation shifting from model performance toward execution responsibility and security structure.[–]nutapongpe·5 か月前返信Trand gold[+]Wizberry·5 か月前
LLM auditing LLM is like RAG solving reasoning. Same blindspot, twice.