Timeline for Are GPT-3.5 series models based on GPT-3?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 4, 2023 at 19:35 | comment | added | nbro | I don't know when I will start reading them. I'll start a project at work that involves these OpenAI models, so I will most likely read these papers (at least to some extent), but I don't know when yet. | |
| Feb 4, 2023 at 17:52 | comment | added | iMad | Thanks for you answer. I went through more or less the same links (that I've included in my question) and I also checked the InstructGPT paper. The different statements from the blog posts are really confusing especially when it comes to different number of parameters... Maybe they use the last layers from GPT-3 to fine tune the new models with fewer parameters.. but in this case they need to run the input through GPT-3 everytime and I'm not sure this is a practical way to do... Anyway I'm waiting for your update. ^^ | |
| Feb 4, 2023 at 17:16 | history | answered | nbro | CC BY-SA 4.0 |