I know we can use few shot prompting to give examples of guided/follow up questions. But I also understand that more than 5/6 examples in a few shot prompting is not recommended. In that case, do we actually need to look at fine-tuning and do instruction fine tuning of models. Did anyone work on this type of scenarios. If yes, it will be great to know their experiences and any specific technique followed.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Creating a Q&A chatbot using fine-tunning with RAG | 3 | 1648 | May 25, 2024 | |
| 3.5t fine tuning with RAG based examples in system prompt? | 0 | 982 | October 19, 2023 | |
| Better Results When Dataset Includes RAG examples? Fine-tune already fine tuned model? | 0 | 608 | November 17, 2023 | |
| Why do some problems after fine-tuning the large model not match the answer | 3 | 379 | August 11, 2024 | |
| Knowledge through fine tuning or RAG embedding | 2 | 500 | September 28, 2024 |