Closed-loop self-evolution on LLMs improves reasoning on Knights and Knaves tasks but plateaus short of oracle-supervised levels, with multi-turn revision nearly matching it for large models.
arXiv preprint arXiv:2601.05280 , doi=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
AI is shifting researchers from creators to curators of generated content, risking loss of intellectual ownership and genuine understanding of science.
citing papers explorer
-
On the Generalization Gap in Self-Evolving Language Model Reasoning
Closed-loop self-evolution on LLMs improves reasoning on Knights and Knaves tasks but plateaus short of oracle-supervised levels, with multi-turn revision nearly matching it for large models.