Why are LLMs so bad at being accurate? And why do people use them for research despite them being so inaccurate?

#LLM #AI #Research #Accuracy

https://www.christopher-james-hall.com/blog/llm-accuracy-and-implications-for-research

Just ask chat? Why are LLMs so inaccurate? — Platform Journalism

AI is a big deal at the moment. Grants are given to research that involves AI. Job ads are asking for AI knowledge or skills. And the first step in learning for many undergraduates, fresh out of high school, is to “just ask chat”. What’s more, the newest version of ChatGPT apparently has “ new PhD a

Platform Journalism

@Platform_Journalism This ain't rocket science. Current generation #LLMs have no semantic layer: they have no model of the world, and no concept of truth. When they are right, it's entirely by accident.

They generate sequences of tokens. The sequences of tokens they generate are probable, given the statistical distribution of similar sequences in texts they have ingested. But there is no concept of meaning there.

People use them in research because they also don't care about truth.