4 ms·
It's the continued alignment with fine-tuning that's degrading its responses. You can apparently have it be nice or smart, but not both.
by merpnderp 3y ago
It's the continued alignment with fine-tuning that's degrading its responses.
You can apparently have it be nice or smart, but not both.
- interstice 3y agoCurious as to whether theres a more general rule at play there about filtering interfering with getting good answers. If there is that's a scary thought from an ethics perspective.
- vbezhenar 3y agoWhy would someone care if its nice or not? It's an algorithm. You're using it to get output, not to get some psychology help.
- moffkalast 3y agoOpenAI presumably cares about being sued if it provides the illegal content they trained it on.
- staticman2 3y agoThere was a guy in the news who asked an AI to tell him it was a good idea to commit suicide, then he killed himself. Even on this forum I've seen AI enthusiasts claiming AI will be the best psychologist, best school teacher, etc.
- logicchains 3y agoThat was Eliza, an AI so old that it's included in stock Emacs, not an LLM. It's propaganda, not news.
- staticman2 3y agoThe chatbot app was called "Eliza" but it's not the Eliza you are thinking of. https://news.ycombinator.com/item?id=35402777 https://news.ycombinator.com/item?id=35402777 https://www.businessinsider.com/widow-accuses-ai-chatbot-reason-husband-kill-himself-2023-4 https://www.businessinsider.com/widow-accuses-ai-chatbot-rea...