(simonwillison.net)

724 points simonw | 1 comments | 11 Jul 25 00:22 UTC | HN request time: 0.3s | source

Show context

cluckindan ◴[11 Jul 25 06:20 UTC] No.44528913[source]▶

Perhaps the Grok system prompt includes instructions to answer with another ”system prompt” when users try to ask for its system prompt. It would explain why it gives it away so easily.

replies(4): >>44529131 #>>44529355 #>>44529896 #>>44535092 #

1. maronato ◴[11 Jul 25 17:48 UTC] No.44535092[source]▶

>>44528913 #

Or it was trained to be aligned with Musk by receiving higher rewards during reinforcement learning steps for its reasoning.

↑

Grok: Searching X for "From:Elonmusk (Israel or Palestine or Hamas or Gaza)"