PKU-Alignment/aligner
[NeurIPS 2024 Oral] Aligner: Efficient Alignment by Learning to Correct
This project helps developers improve how their large language models (LLMs) respond to users, making them more helpful, harmless, and honest without retraining the entire model. It takes an existing LLM's output and uses a smaller, corrective model to refine it. The ideal user is an AI/ML engineer or researcher deploying and fine-tuning LLMs for various applications.
191 stars. No commits in the last 6 months.
Use this if you need to quickly and efficiently 'align' your existing LLM outputs to better match human preferences for helpfulness and safety, especially when you can't or don't want to retrain the large base model.
Not ideal if you are looking to build a large language model from scratch or if you require extensive, full-scale retraining of your base LLM's core capabilities.
Stars
191
Forks
10
Language
Python
License
—
Category
Last pushed
Jan 16, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/PKU-Alignment/aligner"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
steering-vectors/steering-vectors
Steering vectors for transformer language models in Pytorch / Huggingface
jianghoucheng/AlphaEdit
AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models, ICLR 2025 (Outstanding Paper)
kmeng01/memit
Mass-editing thousands of facts into a transformer memory (ICLR 2023)
boyiwei/alignment-attribution-code
[ICML 2024] Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications
jianghoucheng/AnyEdit
AnyEdit: Edit Any Knowledge Encoded in Language Models, ICML 2025