Internal consistency and self-feedback in large language models: A survey

X Liang, S Song, Z Zheng, H Wang, Q Yu, X Li… - arXiv preprint arXiv …, 2024 - arxiv.org
Large language models (LLMs) often exhibit deficient reasoning or generate hallucinations.
To address these, studies prefixed with" Self-" such as Self-Consistency, Self-Improve, and …

Plug-and-play grounding of reasoning in multimodal large language models

J Chen, Y Liu, D Li, X An, W Deng, Z Feng… - arXiv preprint arXiv …, 2024 - arxiv.org
The rise of Multimodal Large Language Models (MLLMs), renowned for their advanced
instruction-following and reasoning capabilities, has significantly propelled the field of visual …

Literature Review of AI Hallucination Research Since the Advent of ChatGPT: Focusing on Papers from arXiv

DM Park, HJ Lee - Informatization Policy, 2024 - koreascience.kr
Hallucination is a significant barrier to the utilization of large-scale language models or
multimodal models. In this study, we collected 654 computer science papers with" …

Super (ficial)-alignment: Strong models may deceive weak models in weak-to-strong generalization

W Yang, S Shen, G Shen, W Yao, Y Liu, Z Gong… - arXiv preprint arXiv …, 2024 - arxiv.org
Superalignment, where humans act as weak supervisors for superhuman models, has
become a crucial problem with the rapid development of Large Language Models (LLMs) …

From imitation to introspection: Probing self-consciousness in language models

S Chen, S Yu, S Zhao, C Lu - arXiv preprint arXiv:2410.18819, 2024 - arxiv.org
Self-consciousness, the introspection of one's existence and thoughts, represents a high-
level cognitive process. As language models advance at an unprecedented pace, a critical …

Unified Active Retrieval for Retrieval Augmented Generation

Q Cheng, X Li, S Li, Q Zhu, Z Yin, Y Shao, L Li… - arXiv preprint arXiv …, 2024 - arxiv.org
In Retrieval-Augmented Generation (RAG), retrieval is not always helpful and applying it to
every instruction is sub-optimal. Therefore, determining whether to retrieve is crucial for …

H2RSVLM: Towards Helpful and Honest Remote Sensing Large Vision Language Model

C Pang, J Wu, J Li, Y Liu, J Sun, W Li, X Weng… - arXiv preprint arXiv …, 2024 - arxiv.org
The generic large Vision-Language Models (VLMs) is rapidly developing, but still perform
poorly in Remote Sensing (RS) domain, which is due to the unique and specialized nature …

Utilize the Flow before Stepping into the Same River Twice: Certainty Represented Knowledge Flow for Refusal-Aware Instruction Tuning

R Zhu, Z Ma, J Wu, J Gao, J Wang, D Lin… - arXiv preprint arXiv …, 2024 - arxiv.org
Refusal-Aware Instruction Tuning (RAIT) enables Large Language Models (LLMs) to refuse
to answer unknown questions. By modifying responses of unknown questions in the training …

Defining Boundaries: A Spectrum of Task Feasibility for Large Language Models

W Zhang, Z Xu, H Cai - arXiv preprint arXiv:2408.05873, 2024 - arxiv.org
Large language models (LLMs) have shown remarkable performance in various tasks but
often fail to handle queries that exceed their knowledge and capabilities, leading to incorrect …

RoleBreak: Character Hallucination as a Jailbreak Attack in Role-Playing Systems

Y Tang, B Wang, X Wang, D Zhao, J Liu… - arXiv preprint arXiv …, 2024 - arxiv.org
Role-playing systems powered by large language models (LLMs) have become increasingly
influential in emotional communication applications. However, these systems are …