Proposes compiling preference pairs into readable natural-language specifications for inference-time LLM alignment, claiming outperformance over DPO on dense-preference domains.
arXiv preprint arXiv:2201.08531 , year=
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2representative citing papers
UltraChat supplies 1.5 million high-quality multi-turn dialogues that, when used to fine-tune LLaMA, produce UltraLLaMA, which outperforms prior open-source chat models including Vicuna.
citing papers explorer
-
Towards Spec Learning: Inference-Time Alignment from Preference Pairs
Proposes compiling preference pairs into readable natural-language specifications for inference-time LLM alignment, claiming outperformance over DPO on dense-preference domains.
-
Enhancing Chat Language Models by Scaling High-quality Instructional Conversations
UltraChat supplies 1.5 million high-quality multi-turn dialogues that, when used to fine-tune LLaMA, produce UltraLLaMA, which outperforms prior open-source chat models including Vicuna.