An analysis of DeepSeek's R1-Zero and R1

Re:An analysis of DeepSeek's R1-Zero and R1 by jvanderbot on Wednesday January 29, 2025 @02:05PM (#42869573)

Really? Isn't that the point of RL used in the way R1 did?

Provide a cost function (vs labels) and have it argue itself to greatness as measured by that cost function?

I believe that's what GP meant by "respond", not telling GPT they were wrong.

[ Parent | on HN ]

Re:An analysis of DeepSeek's R1-Zero and R1 by daveguy on Wednesday January 29, 2025 @03:40PM (#42870870)

That is still inference. It is using a model generated from the RL process. The RL process is what used the cost function to add another model layer. Any online/continual learning would have to be performed by a different algorithm than classical LLM or RL. You can think of RL as a revision, but it still happens offline. Continual learning is still very difficult.

[ Parent | on HN ]

Re:An analysis of DeepSeek's R1-Zero and R1 by jvanderbot on Wednesday January 29, 2025 @04:25PM (#42871373)

Yes, that makes sense. We're both talking about offline learning.

[ Parent | on HN ]