Hugging Face's logo

RLHFlow
university
RLHFlow
RLHFlow
Activity Feed

AI & ML interests

Workflow of Reinforcement Learning from Human Feedback (RLHF). Blog: https://rlhflow.github.io/

Recent Activity
Chenlu123  submitted a paper 3 days ago
Adaptive Layerwise Perturbation: Unifying Off-Policy Corrections for LLM RL
baohao  submitted a paper about 2 months ago
Self-Hinting Language Models Enhance Reinforcement Learning
baohao  updated a collection 5 months ago
Reinforce-Ada
View all activity

Papers

Reinforce-Ada: An Adaptive Sampling Framework for Reinforce-Style LLM Training

View all Papers

RLHFlow 's collections 12

RLHFLow Reward Models
Reward models trained by RLHFlow codebase (https://github.com/RLHFlow/RLHF-Reward-Modeling/)
Standard-format-preference-dataset
We collect the open-source datasets and process them into the standard format.
RM-Bradley-Terry
We train the reward model as the maximum likelihood estimation of the Bradley-Terry model.
Online RLHF
Datasets, code, and models for online RLHF (i.e., iterative DPO)
SFT Models
We train a series of SFT models on the high-quality SFT dataset of RLHFlow for research purpose.
Standard-format-preference-dataset
We collect the open-source datasets and process them into the standard format.
RM-Bradley-Terry
We train the reward model as the maximum likelihood estimation of the Bradley-Terry model.
Online RLHF
Datasets, code, and models for online RLHF (i.e., iterative DPO)
RLHFLow Reward Models
Reward models trained by RLHFlow codebase (https://github.com/RLHFlow/RLHF-Reward-Modeling/)
SFT Models
We train a series of SFT models on the high-quality SFT dataset of RLHFlow for research purpose.