Answer by Mike Li for Have there been attempts to train LLMs using generated...
It has been done before by modifying the TriviaQA dataset in the paper: Alignment for Honesty (Yang et al.). They try multiple techniques including outputting the confidence score in the response.They...
View ArticleHave there been attempts to train LLMs using generated "I don't know" data to...
Correct me if I am wrong, but I assume that hallucinations can be primarily blamed on the training data underrepresenting responses that specifically state that someone doesn't know a specific fact. On...
View ArticleAnswer by Tanveer for Have there been attempts to train LLMs using generated...
Unfortunately no, no such work has been done (to my knowledge).Let's go back to your question. Hallucinations come from a prediction that does not meet the historical representation of actual events....
View ArticleAnswer by Franck Dernoncourt for Have there been attempts to train LLMs using...
It would be fairly easy to generate a big set of training data that's basically countless of questions for which there are probably no known answers with a simple "I don't know" response.It may...
View Article
![[QLD, NSW, VIC] Samsung 65](http://files.ozbargain.com.au/n/31/978831l.jpg?h=8d64f0c6)
