home/categories/machine-learning/orchestra-research-ai-research-skills-06-post-training-simpo-skill-md
machine-learningdata-ai

simpo-training

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.

Orchestra-Research
maintainer
Orchestra-Research
Updated 11/20/2025
Stars
6563
Forks
515
quick start

Installation and usage

Simple Preference Optimization for LLM alignment. Reference-free alternative to DPO with better performance (+6.4 points on AlpacaEval 2.0). No reference model needed, more efficient than DPO. Use for preference alignment when want simpler, faster training than DPO/PPO.

Installation
$ install --globalskills.sh
Usage

Once installed, you can use this skill by running the following command in your terminal:

skills use simpo-training