yep. Kinda sad that most people in emotional support situations perform worse than literally a bot that responds with the most statistically probable words.
My guess, the subject matter likely triggers a safety protocol. How often does the bot cite sources in that way? It’s not a conversational reply that’s trying to emulate a human.
yep. Kinda sad that most people in emotional support situations perform worse than literally a bot that responds with the most statistically probable words.
My guess, the subject matter likely triggers a safety protocol. How often does the bot cite sources in that way? It’s not a conversational reply that’s trying to emulate a human.
All the time