rlhf

9 progetti condividono questo topic GitHub

Le linee collegano membri che sono misurabilmente correlati tra loro. La dimensione dei punti riflette le stelle.

🧬 Membri

alignment-handbook

Robust recipes to align language models with human and AI preferences

An Efficient and User-Friendly Scaling Library for Reinforcement Learning with Large Language Models

Textbook on reinforcement learning from human feedback

[NeurIPS 2023] ImageReward: Learning and Evaluating Human Preferences for Text-to-image Generation

WebGLM: An Efficient Web-enhanced Question Answering System (KDD 2023)

RLHF-Reward-Modeling

Recipes to train reward model for RLHF.

[NeurIPS 2024] SimPO: Simple Preference Optimization with a Reference-Free Reward

The official implementation of Self-Play Preference Optimization (SPPO)

Code For Adaptive-Boundary-Clipping GRPO. arxiv.org/pdf/2601.03895

🔗 Famiglie affini

alignment 14 transformers 55 agentic 69 generative-model 45 llama3 7 grpo 19

Misurato dai temi di GitHub condivisi da entrambi i progetti, ponderato in base a quanto è raro ciascun tema.