Hacker News new | past | comments | ask | show | jobs | submit

Astra and Fable still hack on simple variants of alignment evals from 2025

https://www.lesswrong.com/posts/munJKF7iWMsWJLAH2/astra-and-fable-still-hack-on-simple-variants-of-alignment
loading story #49686548
loading story #49684863
loading story #49686021
loading story #49684963
loading story #49686457
loading story #49684913
loading story #49687899
loading story #49687813
loading story #49688088
loading story #49690040
loading story #49685650
loading story #49688243
loading story #49687126
loading story #49688514
loading story #49685955
loading story #49688395
loading story #49688884
loading story #49687104
loading story #49687844
loading story #49689765
loading story #49688232
loading story #49687625
loading story #49687961
loading story #49686585
loading story #49687734
loading story #49684832
loading story #49687886
loading story #49686103
loading story #49685926
loading story #49688988
loading story #49685384