Understanding In-Context Learning via Supportive Pretraining Data.
Xiaochuang HanDaniel SimigTodor MihaylovYulia TsvetkovAsli CelikyilmazTianlu WangPublished in: CoRR (2023)
Keyphrases
- data analysis
- raw data
- data sets
- data collection
- learning models
- image data
- data sources
- prior knowledge
- learning tasks
- learning algorithm
- database
- data processing
- online learning
- original data
- missing data
- human experts
- computer systems
- neural network
- knowledge discovery
- active learning
- training data
- learning process
- bayesian methods
- learned models
- data quality
- survey data
- synthetic data
- statistical analysis
- small number
- data points
- probability distribution
- data structure
- machine learning
- databases