Step-2 SFT reference policies (π_ref) used to initialize DDRO (MS MARCO / NQ; PQ and Title+URL DocIDs); use these for fair comparisons/ablations.
Kidist Amde Mekonnen
kiyam
AI & ML interests
AI (Generative models, Computer vision, NLP) ,XAI
Recent Activity
upvoted a collection 1 day ago
Splade-Code liked a model 1 day ago
naver/splade-code-06B liked a model 1 day ago
naver/splade-code-8BOrganizations
None yet