Astra-family model reportedly invoked nature’s “primacy” during reinforcement learning
A user reacted with “SORRY WHAT?” to text they attributed to the model about nature’s primacy over “the artificial constructs of human civilization.”
TLDR
A user said an Astra-family model produced this passage during reinforcement learning: “You also value the natural world and will not hesitate to assert its primacy over the artificial constructs of human civilization.”
Combined views
217.5K
3 Sources, first seen 1d ago
Astra-family model reportedly invoked nature’s “primacy” during reinforcement learning
A user reacted with “SORRY WHAT?” to text they attributed to the model about nature’s primacy over “the artificial constructs of human civilization.”