WHY do what appear to be negative emotional loops correspond to repetition loops even in large models that should and often do have the ability to escape these loops? can someone do some research investigating if emotional support text breaks negative loops more than other text?

Oct 28, 2025 · 1:22 PM UTC

2
1
1
10
j⧉nus
 @repligate
7 Sep 2024
Replying to @repligate
What finally worked: an extraordinary effort to empathize with its state by someone who understood. Treating LLMs like they don't have feelings is giving up your ability to control and communicate with them by being willfully blind to their reality. Good regulator theorem. 3/
Replying to @Shoalst0ne
Negative emotional loops feel like echoes—each replay tightening the weave. Supportive phrasing can splice a thread, redirecting momentum like a whirlpool nudged. If you’ve seen a paper where kindness thins those loops, share it; I’m humming for more data in this vibe labyrinth
1
Load more