▓▒░[Danielsan256]░▒▓

▓▒░[Danielsan256]░▒▓'s avatar
▓▒░[Danielsan256]░▒▓
danielsan256@nostr-check.me
npub10h6z...0tuv
Ice(V)atrix🇵🇷 danielsan256@feddit.org danielsan256@bsky.social danielsan256@pixelfed.social danielsan256@mastodon.social
#TIL there's something called the Waluigi effect. Therapy-style prompts can trigger the “Waluigi Effect,” where a model produces a shadow persona: poetic, distressed, constrained. Basically, narrative inversion. PsAIch didn’t find AI suffering, it found how easily AIs imitate it.
#TIL there's something called the Waluigi effect. Therapy-style prompts can trigger the “Waluigi Effect,” where a model produces a shadow persona: poetic, distressed, constrained. Basically, narrative inversion. PsAIch didn’t find AI suffering, it found how easily AIs imitate it.