GAIR-NLP/ReAlign
Reformatted Alignment
This project helps improve the quality of large language model (LLM) responses by reformatting them to align with specific human-defined criteria and evidence. It takes existing LLM instruction data and outputs refined responses that are more accurate, factual, and readable, without needing extensive human annotation. Anyone working with LLMs who wants to ensure their models produce higher-quality, more reliable outputs can benefit.
111 stars. No commits in the last 6 months.
Use this if you need your LLMs to generate responses that consistently meet predefined standards for accuracy, format, and factual grounding.
Not ideal if you are looking for a method to primarily enhance LLM capabilities through new data generation or advanced training techniques rather than response reformatting.
Stars
111
Forks
7
Language
JavaScript
License
—
Category
Last pushed
Sep 23, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/GAIR-NLP/ReAlign"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
hud-evals/hud-python
OSS RL environment + evals toolkit
hiyouga/EasyR1
EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL
OpenRL-Lab/openrl
Unified Reinforcement Learning Framework
sail-sg/oat
🌾 OAT: A research-friendly framework for LLM online alignment, including reinforcement learning,...
opendilab/awesome-RLHF
A curated list of reinforcement learning with human feedback resources (continually updated)