3 ms·
In the "sparks of AGI" paper, authors noted that the unicorn shape degrees as more "alignment" is injected to to. If openai adjust the model (say by training mo
by qumpis 4y ago
In the "sparks of AGI" paper, authors noted that the unicorn shape degrees as more "alignment" is injected to to. If openai adjust the model (say by training more), the picture should reflect it. If they make the model be more "aligned", it should reflect as well.
So I'd guess the answer is the former.
- imdsm 3y agoThe former yes. The expectation is that updates will start to become more and more iterative over time. For this reason, I record the specific version that is pegged to the generic `gpt-4` model.