一作ICLR 2026Poster· Reinforcement Learning / Everything Else
When Data is the Algorithm: A Systematic Study and Curation of Preference Optimization Datasets
Aladin Djuhera · Farhan Ahmed · Swanand Kadhe · Syed Zawad · Heiko Ludwig · Holger Boche
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Aladin Djuhera · Farhan Ahmed · Swanand Kadhe · Syed Zawad · Heiko Ludwig · Holger Boche
Aladin Djuhera · Swanand Ravindra Kadhe · Farhan Ahmed · Syed Zawad · Holger Boche
Fine-tuning large language models (LLMs) is a common practice to adapt generalist models to specialized domains. However, recent studies show that fine-tuning can erode safety alignment, causing LLMs to respond to harmful or unethical prompts. Many methods to realign safety have been proposed, bu...