This was an interesting raising-awareness project.
And the article says they didn’t let the chatbot generate its own responses (and therefore produce LLM hallucinations) but rather used an LLM in the background to categorize user’s question and return an answer from said category.
This is what the hell we need. I’m sure someone could use an LLM to make natural language processing easy. Until then, I’ll keep listening to music whenever I tell Google to turn the lights on.
This was an interesting raising-awareness project.
And the article says they didn’t let the chatbot generate its own responses (and therefore produce LLM hallucinations) but rather used an LLM in the background to categorize user’s question and return an answer from said category.
This is what the hell we need. I’m sure someone could use an LLM to make natural language processing easy. Until then, I’ll keep listening to music whenever I tell Google to turn the lights on.