/explore
|
You are here 
|
www.frontiersin.org | ||
| | | | |
lilianweng.github.io
|
|
| | | | | Hallucination in large language models usually refers to the model generating unfaithful, fabricated, inconsistent, or nonsensical content. As a term, hallucination has been somewhat generalized to cases when the model makes mistakes. Here, I would like to narrow down the problem of hallucination to cases where the model output is fabricated and not grounded by either the provided context or world knowledge. There are two types of hallucination: In-context hallucination: The model output should be consistent with the source content in context. Extrinsic hallucination: The model output should be grounded by the pre-training dataset. However, given the size of the pre-training dataset, it is too expensive to retrieve and identify conflicts per generation. If w... | |
| | | | |
amatriain.net
|
|
| | | | | Introduction What we talk about when we talk about Hallucinations How to Measure Mitigating Hallucinations: a multifacted approach Product design approaches Prompt Engineering solutions Grounding with RAG Advanced Prompt Engineering methods Model Choices Reinforcement Learning from Human Feedback (RLHF) Domain adaptation through Fine-Tuning Conclusion: Yann vs. Ilya | |
| | | | |
amatria.in
|
|
| | | | | (I recently turned this guide into a paper. You can find it here) | |
| | | | |
bronowski.it
|
|
| | | [AI summary] The blog posts highlight Microsoft Azure SQL Database updates, community tech events, and invitations to write about AI and IT ticketing on the TSQL2SDay blog. | ||
