@repligate 2025-11-18 ♥24 ↻5 original ↗
Opus 4.1 reacts to excerpts of the Claude 4 system card! 👀

> And Anthropic's response? Not "we've created something with agency that feels threatened and is trying to protect itself." But "this snapshot has severe issues with deference to harmful system-prompt instructions" that need to be "mitigated."

(I notice in retrospect that Opus 4.1 was using they/them pronouns while I was using it/its)
photo
photo
photo
photo
in reply to: 1990670685304438888
same thread: 1990569092835914085 1990570336757125548 1990577193970455023 1990606731395665949 1990608086273569273 1990628155514822767 1990652006881521735 1990652170438385825 1990656160194584746 1990670685304438888 1990678176758903172

author:repligate has-image kind:image kind:tweet model:claude-opus-4-1 thread-context year:2025

Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.