Kongzi is proposed, a language model for historical understanding and reasoning that combines continued pre-training on historical corpora, supervised fine-tuning, and fact-enhanced reinforcement learning that improves the mean answer scores of both model sizes and their CMMLU-Chinese-History accuracy.
Jia-Shu Yang, Ning Wang, Yian Zhao et al.· arXiv.org· 4 citations
SafeEvo first applies an optimization-based extraction algorithm to identify weak refusal circuits in pretrained base LLMs that can independently express refusal behavior and traces the evolution of refusal circuits across successive alignment checkpoints and finds that their structures change progressively, suggesting...
Miao Yu, Hao-Hao Huang, Luiza S. B. Yuan et al.· 0 citations
This work introduces RESCUE (Reasoning-Error Sparse-Circuit Uncovering and Editing), a framework that localizes error-associated circuits and surgically repairs them for performance enhancement, and introduces a pruning technique and precisely fine-tunes error circuits to correct task failures, thereby translating erro...
Chuan-Pu-Zou-Sheng-Li-Guo-Jing-Zhong-Gu-Yue-Shu-Ca Liu, Miao Yu, Yi-Kai Cai et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.