Re:Reasoning Is Not Model Improvement
by Terr_ on Thursday October 23, 2025 @07:21PM
(#45688727)
Some nice charts here [0], which IMO means LLMs are getting very good at guessing answers to certain arithmetic operations, but they don't actually perform it in a logical fashion.
[0] https://www.mindprison.cc/p/why-llms-dont-ask-for-calculator...