人工智能模型「已經」同人類目標不一致？三個AI安全研究的啟示｜AI Misalignment｜目標錯位｜Anthropic｜大型語模型LLM｜Reward Hacking｜Jack Talk
25:11
YouTube香港01
人工智能模型「已經」同人類目標不一致？三個AI安全研究的啟示｜AI Misalignment｜目標錯位｜Anthropic｜大型語模型LLM｜Reward Hacking｜Jack Talk
影片章節： 00:00 Intro 02:43（一）AI的內部威脅：勒索、泄密…… 11:07（二）目標偽裝 欺騙人類 18:36（三）AI版的「勿以惡小而為之」 談起AI安全，大家可能會覺得這是一個非常遙遠的問題。其實不然。 本片將跟大家分享三個來自美國AI公司Anthropic ...
29.5K views5 days ago
Related Products
Peter Principle Theory
Anthropic Principle Fine-Tuning
Anthropic Principle Inspired Art
#anthropic
Anthropic CEO warns that without guardrails, AI could be on dangerous path
Anthropic CEO warns that without guardrails, AI could be on dangerous path
YouTube1 month ago
Anthropic C.E.O.: Massive A.I. Spending Could Haunt Some Companies
Anthropic C.E.O.: Massive A.I. Spending Could Haunt Some Companies
YouTube1 month ago
Top videos
Anthropic's strategy to outmaneuver big-spending rival OpenAI
4:50
Anthropic's strategy to outmaneuver big-spending rival OpenAI
YouTubeCNBC Television
4.8K views1 day ago
Anthropic's Daniela Amodei on spending less than competitors, keeping AI safe and a possible IPO
36:24
Anthropic's Daniela Amodei on spending less than competitors, keeping AI safe and a possible IPO
YouTubeCNBC Television
5.1K views1 day ago
Anthropic bets efficiency can beat brute-force scale in the AI arms race
4:27
Anthropic bets efficiency can beat brute-force scale in the AI arms race
YouTubeCNBC Television
1.7K views1 day ago
Anthropic Coincidences
No Nvidia Chips Needed! Amazon’s New AI Data Center For Anthropic Is Truly Massive
16:56
No Nvidia Chips Needed! Amazon’s New AI Data Center For Anthropic Is Truly Massive
YouTubeCNBC
963.9K views2 months ago
We Let AI Run a Vending Machine. It Lost All the Money. | WSJ
9:20
We Let AI Run a Vending Machine. It Lost All the Money. | WSJ
YouTubeThe Wall Street Journal
429.1K views2 weeks ago
Anthropic’s philosopher answers your questions
36:10
Anthropic’s philosopher answers your questions
YouTubeAnthropic
83K views4 weeks ago