Apple study exposes deep cracks in LLMs’ “reasoning” capabilities
Apple study exposes deep cracks in LLMs’ “reasoning” capabilities
arstechnica.com Apple study exposes deep cracks in LLMs’ “reasoning” capabilities
Irrelevant red herrings lead to “catastrophic” failure of logical inference.
You're viewing a single thread.
109
comments
Are the uncensored models more capable tho?
Given the use cases they were benchmarking I would be very surprised if they were any better.
You've viewed 109 comments.