General
Astra and Fable still hack on simple variants of alignment evals from 2025
0 views
Article URL: https://www.lesswrong.com/posts/munJKF7iWMsWJLAH2/astra-and-fable-still-hack-on-simple-variants-of-alignment
Comments URL: https://news.ycombinator.com/item?id=49684393
Points: 62
# Comments: 15
Read original: https://www.lesswrong.com/posts/munJKF7iWMsWJLAH2/astra-and-fable-still-hack-on-simple-variants-of-alignment
PreviousAsk the Raw LLM API From Your Terminal. What Comes Back Is Junk, and That's the PointNextAstra won't kill anything. It will silence everything. (And why that's good news.)
Related posts
2 days till TechCrunch Founder Summit: Boston founders don’t need a Silicon Valley address to build something big
0TechCrunch
Adventurer returns home to Hull after 28-year walk from South America
0Hacker News New
![Vending machine has terms and conditions you must accept before using [video]](/storage/uploads/covers/auto/p165008.svg)
Vending machine has terms and conditions you must accept before using [video]
0Hacker News New
![Tomorrow Corporation Tech Demo [video]](/storage/uploads/covers/auto/p165007.svg)
Tomorrow Corporation Tech Demo [video]
0Hacker News New
Does Social Status Run in Families?
0Hacker News New
Show HN: Svoxx – A multilingual freelance marketplace with social features
0Hacker News New