Omni-DPO: A Dual-Perspective Paradigm for Dynamic Preference Learning of LLMs
Paper
•
2506.10054
•
Published
•
3
Uni-DPO
[ICLR 2026] Official repository of "Uni-DPO: A Unified Paradigm for Dynamic Preference Optimization of LLMs". Repo: https://github.com/pspdada/Uni-DPO
-
-
psp-dada/Uni-DPO
Preview • Updated • 28 • 1 -
psp-dada/Qwen2.5-7B-Uni-DPO
Text Generation • 8B • Updated • 14 • 1 -
psp-dada/Llama-3-8B-Instruct-Uni-DPO-v2-GPT-4o
Text Generation • 8B • Updated • 7 • 1 -
psp-dada/Llama-3-8B-Instruct-Uni-DPO-v2-ArmoRM
Text Generation • 8B • Updated • 28 • 1 -
psp-dada/Llama-3-8B-Base-SFT-Uni-DPO
Text Generation • 8B • Updated • 11 • 1 -
psp-dada/Llama-3-8B-Base-SFT-Uni-DPO-v2-Qwen
Text Generation • 8B • Updated • 29 • 1 -
psp-dada/Gemma2-9B-IT-Uni-DPO
Text Generation • 9B • Updated • 13 • 1 -
psp-dada/Llama-3-8B-Base-SFT-Uni-DPO-v2-GPT-4
Text Generation • 8B • Updated • 15 • 1 -
psp-dada/Llama-3-8B-Instruct-Uni-DPO
Text Generation • 8B • Updated • 15 • 1 -
psp-dada/Qwen2.5-Math-7B-Uni-DPO
Text Generation • 8B • Updated • 17 • 1