Is Chain-of-Thought Reasoning of LLMs a Mirage? A Data Distribution Lens (2025)arxiv.org1 points by SomaticPirate 8 hours ago
SomaticPirate 7 hours agoA useful reminder that just because an LLM "appears" to be thinking and reasoning that it likely isn't. If it hasn't seen something similar or "in distribution" before, then it typically doesn't perform well.
A useful reminder that just because an LLM "appears" to be thinking and reasoning that it likely isn't. If it hasn't seen something similar or "in distribution" before, then it typically doesn't perform well.