Skip to main content
Free forever · no paywall · no ads
Request a course
→
en
vi
claudem
y
.org
Tracks
Library
By use case
Skills
Search courses, lessons…
⌘K
en
vi
Home
YouTube Library
Anthropic — Research Papers
Course · YouTube Library
Anthropic — Research Papers
Anthropic
19 lessons
11h 26m
Giải thích cơ chế và an toàn AI
1
What is interpretability?
advanced
4m
AI personality and alignment methods
2
What should an AI's personality be?
advanced
38m
Scaling LLM interpretability
3
Scaling interpretability
advanced
53m
AI policy, safety, interpretability, future
4
AI, policy, and the weird sci-fi future with Anthropic’s Jack Clark
intermediate
38m
AI usage analysis for safety & impact
5
What do people use AI models for?
intermediate
47m
LLM alignment faking and safety implications
6
Alignment faking in large language models
advanced
1h 30m
Challenges in AI alignment and interpretability
7
How difficult is AI alignment? | Anthropic Research Salon
advanced
28m
AI safety and jailbreak prevention
8
Defending against AI jailbreaks
intermediate
1h 15m
AI safety, alignment, and control
9
Controlling powerful AI
advanced
51m
AI interpretability and internal thought processes
10
Tracing the thoughts of a large language model
intermediate
3m
AI consciousness and ethics
11
Could AI models be conscious?
intermediate
44m
AI ethics and societal impact
12
The Societal Impacts of AI
intermediate
8m
AI emotional support use safety research
13
Affective Use of AI
beginner
12m
understanding LLM internal mechanisms
14
Interpretability: Understanding how AI models think
advanced
59m
AI cybercrime and future safety threats
15
Threat Intelligence: How Anthropic stops AI cybercrime
intermediate
37m
Reward hacking and AI alignment
16
What is Al "reward hacking"—and why do we worry about it?
advanced
52m
AI ethics, identity, and welfare
17
Anthropic’s philosopher answers your questions
intermediate
36m
AI model alignment and safety
18
What is sycophancy in AI models?
intermediate
6m
AI functional emotions & interpretability
19
When AIs act emotional
intermediate
5m
💬
Feedback / Report
Spotted an issue or have an improvement idea?
→