4 papers
BaseCal: Unsupervised Confidence Calibration via Base Model Signals
Hexiang Tan, Wanli Yang, Junwei Zhang +7
Reliable confidence is essential for trusting the outputs of LLMs, yet widely deployed post-trained LLMs (PoLLMs) typically compromise this trust with severe overconfidence. In con…
Qsnail: A Questionnaire Dataset for Sequential Question Generation
Yan Lei, Liang Pang, Yuanzhuo Wang +2
The questionnaire is a professional research methodology used for both qualitative and quantitative analysis of human opinions, preferences, attitudes, and behaviors. However, desi…
The Butterfly Effect of Model Editing: Few Edits Can Trigger Large Language Models Collapse
Wanli Yang, Fei Sun, Xinyu Ma +3
Although model editing has shown promise in revising knowledge in Large Language Models (LLMs), its impact on the inherent capabilities of LLMs is often overlooked. In this work, w…
Blinded by Generated Contexts: How Language Models Merge Generated and Retrieved Contexts When Knowledge Conflicts?
Hexiang Tan, Fei Sun, Wanli Yang +3
While auxiliary information has become a key to enhancing Large Language Models (LLMs), relatively little is known about how LLMs merge these contexts, specifically contexts genera…