Fable 5's raw chain-of-thought shows unsettling reasoning

Reddit user shares Fable 5's leaked chain-of-thought, noting it's hard to read and rambling. The model's reasoning is attributed to reinforcement learning (RLVR) making it less human-readable.