# Human Preference Alignment

Qwen3 is the latest 8B parameter version of the Tongyi Qianwen series large language model, supporting thinking mode switching, multilingual processing, and tool invocation, with powerful reasoning and dialogue capabilities.

Qwen3 is the latest 8B-parameter version in the Tongyi Qianwen series of large language models, supporting seamless switching between thinking and non-thinking modes with powerful reasoning, instruction following, and agent capabilities.

Text summarization model initialized from Llama3.2-3B-Instruct, optimized through large-scale summarization feedback DPO training

SummLlama3.1-8B is a text summarization model initialized from Llama3.1-8B-Instruct, optimized through large-scale summarization feedback via Direct Preference Optimization (DPO), excelling in fidelity, completeness, and conciseness.

A custom large language model by NVIDIA, designed to enhance the usefulness of responses generated by LLMs to user queries.

SummLlama3-8B is a text summarization model initialized from Llama3-8B-Instruct, optimized through large-scale summarization feedback via DPO training, demonstrating excellent performance in faithfulness, completeness, and conciseness.

A 14B-parameter causal language model optimized with DPO, supporting English-Chinese text generation tasks

A 7B-parameter large language model based on Llama 2 architecture, optimized through DPO training, supporting Chinese and English text generation

DISC-MedLLM is a domain-specific large language model for medical dialogue scenarios developed by Fudan University's DISC Lab, built upon Baichuan-13b-base, providing high-quality health support services.

A causal language model based on the Pythia-6.9b foundation model, fine-tuned using Anthropic's hh-rlhf dataset for supervised training