Dangerous Capability Evals: Basis for Frontier Safety | Mary Phuong (Google DeepMind)
Published September 10, 2024
At a glance
Views
28,552
Likes
207
Comments
4
Performance vs. channel
All-time outlier score
Engagement rate
More features coming
We're working on new analytics and tools. Stay tuned.
About
Mary Phuong (Google DeepMind) on how dangerous capability evaluations work, and why they are the technical basis for frontier AI safety frameworks. In this FAR.AI talk, Mary Phuong of Google DeepMind explains how labs use dangerous capability evaluations to anticipate severe risks from frontier models before they arri…
- Published
- September 10, 2024
- Made for kids
- No
More features coming
We're working on new analytics and tools. Stay tuned.
How Dangerous Capability Evals: Basis for Frontier Safety | Mary Phuong (Google DeepMind) is doing
It has 28.5K views in total. It ranks #14,595,940 among tracked YouTube videos by views.
Key facts about Dangerous Capability Evals: Basis for Frontier Safety | Mary Phuong (Google DeepMind)
- Total views
- 28.5K
- Total likes
- 207
- Total comments
- 4
- Published
- September 10, 2024
- Global rank by views
- #14,595,940
Frequently asked questions
How many views does Dangerous Capability Evals: Basis for Frontier Safety | Mary Phuong (Google DeepMind) have?
Dangerous Capability Evals: Basis for Frontier Safety | Mary Phuong (Google DeepMind) has 28,552 views, 207 likes and 4 comments, ranking #14,595,940 among tracked YouTube videos by views.
How accurate are these YouTube video analytics?
View, like, and comment counts come from YouTube's public API and refresh regularly. Growth and revenue figures are estimates built from tracked historical data, so they can differ slightly from what the creator sees privately in YouTube Studio.
What is the outlier score?
How this video's views compare to the channel's usual Shorts or long-form video. A 2x score means double the usual views for that channel, 0.5x means half.
What is engagement rate and how is it calculated?
Likes and comments per view, compared to this channel's typical engagement rate — shown as above or below what's usual for the channel.
How is the revenue estimate calculated?
It's a rough public estimate based on a $0.7–$2.0 CPM (revenue per 1,000 views) applied to view count — not actual creator earnings, which depend on ad rates, audience geography, and monetization settings YouTube doesn't expose publicly.
Is this YouTube video analytics checker free?
Yes. You get 10 free channel and video analytics lookups per 24 hours without an account — logging in, which is also free, removes that limit entirely.