MARVEL: Modular Abstention for Reliable and Versatile Expert LLMs
A modular abstention framework for reliable expert LLMs that enables selective abstention from uncertain questions.
Do Language Models Mirror Human Confidence? Exploring Psychological Insights to Address Overconfidence in LLMs
Exploring psychological insights to address overconfidence in LLMs by comparing with human confidence patterns.
Chenjun Xu*
•
•
1 min read
AutoScale-Automatic Prediction of Compute-optimal Data Composition for Training LLMs
Automatic prediction of compute-optimal data composition for efficient LLM training.
Feiyang Kang
•
•
1 min read
Know Your Limits: A Survey of Abstention in Large Language Models
A comprehensive survey of abstention mechanisms in large language models, covering theory, implementation, and evaluation.
Characterizing LLM Abstention Behavior in Science QA with Context Perturbations
Characterizing LLM abstention behavior in science QA with context perturbations.