Jie Lu

Detecting hallucinations in Retrieval-Augmented Generation remains a challenge. Prior approaches attribute hallucinations to a binary conflict between internal knowledge stored in FFNs and the retrieved context. However, this perspective is incomplete, failing to account for the impact of other c...

Hallucination detection is crucial for large language models (LLMs), as hallucinated content creates significant barriers in applications requiring factual accuracy. Current detection methods mainly depend on internal signals like uncertainty and self-consistency checks, using the model’s pre-tra...

Implicit feedback, such as clicks and browsing behaviors, is ubiquitous in recommender systems as a proxy for user preferences. However, these signals are inherently noisy; interactions such as misclicks, unintended views, or unsatisfactory purchases often introduce false positive patterns that m...