Reconstructing the benchmark behind Luc Julia's 64% LLM reliability claim github.com 4 points by matthieu_bl 10 days ago · 1 comment Reader PiP Save No comments yet.