AikyamLab/llm-memorization
Understanding the memorization property of Large Language Models using Model Attribution
This project helps AI researchers and developers understand why Large Language Models (LLMs) memorize their training data. By analyzing the internal architecture, specifically attention modules, it shows which parts of an LLM are responsible for memorization versus generalization. It takes an LLM and tokenized memorized data as input, and outputs insights into architectural factors influencing memorization and performance, allowing for more ethical and private LLM deployment.
No commits in the last 6 months.
Use this if you are a machine learning researcher or engineer designing and deploying Large Language Models and need to diagnose and mitigate data memorization to improve privacy and ethical compliance.
Not ideal if you are an end-user simply looking to use an LLM without needing to understand or modify its internal architectural properties related to memorization.
Stars
9
Forks
—
Language
Python
License
MIT
Category
Last pushed
Mar 17, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/AikyamLab/llm-memorization"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
MadryLab/context-cite
Attribute (or cite) statements generated by LLMs back to in-context information.
microsoft/augmented-interpretable-models
Interpretable and efficient predictors using pre-trained language models. Scikit-learn compatible.
Trustworthy-ML-Lab/CB-LLMs
[ICLR 25] A novel framework for building intrinsically interpretable LLMs with...
poloclub/LLM-Attributor
LLM Attributor: Attribute LLM's Generated Text to Training Data
THUDM/LongCite
LongCite: Enabling LLMs to Generate Fine-grained Citations in Long-context QA