A report on base, instruction tuned and chat tuned models for tasks

zeno-ml / zeno-build

Build, evaluate, understand, and fix LLM-based apps

MIT License

484 stars 33 forks source link

A report on base, instruction tuned and chat tuned models for tasks #143

Closed karthik19967829 closed 10 months ago

karthik19967829 commented 1 year ago

Hi Team, Thank you for the great effort, it would be great if a report on base, instruction-tuned, and chat-tuned cases for a specific model like MPT or any other model for tasks can be provided to study the impact of chat and instruction tuning on tasks

neubig commented 1 year ago

Thanks @karthik19967829! I think that'd be quite interesting and we'd welcome contributions. We'd have to decide specifically which tasks to do this for. One potentially related issue is this one: https://github.com/zeno-ml/zeno-build/issues/142