2 ms·
They have, although not datasets: https://openai.com/blog/deep-reinforcement-learning-from-human-preferences/ https://openai.com/blog/deep-reinforcement-learni
by astrange 4y ago
They have, although not datasets:
https://openai.com/blog/deep-reinforcement-learning-from-human-preferences/ https://openai.com/blog/deep-reinforcement-learning-from-hum...
https://openai.com/blog/fine-tuning-gpt-2/ https://openai.com/blog/fine-tuning-gpt-2/
https://openai.com/blog/chatgpt/ https://openai.com/blog/chatgpt/
People like Anthropic have been able to reproduce it.