Large language models (LLMs) have demonstrated remarkable capabilities across a wide
range of tasks in various domains. Despite their impressive performance, they can be …

Direct Preference Optimization (DPO) is a widely used offline preference optimization
algorithm that reparameterizes reward functions in reinforcement learning from human …

Despite their impressive capabilities, large language models (LLMs) have been observed to
generate responses that include inaccurate or fabricated information, a phenomenon …

Large language models are known to hallucinate when faced with unfamiliar queries, but
the underlying mechanism that govern how models hallucinate are not yet fully understood …

Learning from feedback reduces the hallucination of multimodal large language models
(MLLMs) by aligning them with human preferences. While traditional methods rely on labor …

Confidence estimation aiming to evaluate output trustability is crucial for the application of
large language models (LLM), especially the black-box ones. Existing confidence estimation …

Large Vision-Language Models (LVLMs) have made substantial progress by integrating pre-
trained large language models (LLMs) and vision models through instruction tuning. Despite …

The remarkable advancements in Multimodal Large Language Models (MLLMs) have not
rendered them immune to challenges, particularly in the context of handling deceptive …

Large Language Models (LLMs) are notorious for blending fact with fiction and generating
non-factual content, known as hallucinations. To address this challenge, we propose an …

When large language models are aligned via supervised fine-tuning, they may encounter
new factual information that was not acquired through pre-training. It is often conjectured that …