This is a heavily interactive web application, and JavaScript is required. Simple HTML interfaces are possible, but that is not what this is.
Post
Gokul Swamy
gokul.dev
did:plc:7e3hw64shux7ikibrebi6xx5
I think the main difference in terms of interpolation / extrapolation between DPO and RLHF is that the former only guarantees closeness to the reference policy on the training data, while RLHF usually tacks on an on-policy KL penalty. We explored this point in arxiv.org/abs/2406.01462.
2024-11-22T15:38:59.261Z