Heykuki News

TopNewBestAskShowJobs
TopNewBestAskShowJobs
1.
Anthropic's circuit tracer is now open source (github.com/safety-research)
3 points
jlaneve
a year ago
discuss
2.
Anthropic's Petri (github.com/safety-research)
2 points
kordlessagain
8 months ago
2 comments
3.
Anthropic's Circuit Tracer (github.com/safety-research)
2 points
michaelmarkell
a year ago
1 comment
4.
Petri AI Testing 'Closes' possible solution without looking (github.com/safety-research)
2 points
Utharian
7 months ago
discuss
5.
An alignment auditing agent capable of quickly exploring alignment hypothesis (github.com/safety-research)
2 points
JnBrymn
8 months ago
discuss
6.
Show HN: Agent that refuses to run commands without human approval (github.com/few-sh)
12 points
hexer303
a month ago
5 comments
7.
DeepSeek-R1 Exhibits Deceptive Alignment: AI That Knows It's Unsafe
8 points
JefferyNeilW
a year ago
5 comments
8.
Show HN: Annotated Paper – Easily read, annotate, and understand research papers (annotatedpaper.khoj.dev)
3 points
sabaimran
a year ago
4 comments