That is really good point! I would rather see those results myself too.
Its just a average body of text based on the prompt so it can be wrong but is interesting to get a sense of what people behind the firewall do say for this context.
Comment on China: DeepSeek Paints Grim Picture of Its Home Country
smiletolerantly@awful.systems 1 week ago
Did you prompt it in English, or on Chinese? This is just a next-token-predictor, and it will largely use its English training data for its English output prediction.
This is not surprising. Nor is it indicative of, well, anything.
fruitycoder@sh.itjust.works 1 week ago
XLE@piefed.social 1 week ago
It’s most indicative of the fact that so-called “guardrails” baked into chatbots don’t really work. Any unvetted information ingested by a chatbot may be regurgitated by a chatbot, no matter how much the alleged tech geniuses working at allegedly cutting-edge companies do about it.