Extractable Memorization From First Principles

arXiv:2607.12649 · cs.LG, cs.CL · Submitted 2026-07-14 · Read on arXiv

cs.LG, cs.CL

Submitted: 2026-07-14

Updated: 2026-09-25

Terminology

• Exchangeability — 10× in this paper · explained in 2 episodes
• Calibration — 6× in this paper · explained in 7 episodes
• Adam — 4× in this paper · explained in 2 episodes
• Decode — 3× in this paper · explained in 2 episodes
• Generative AI — 3× in this paper · explained in 2 episodes
• Verification — 3× in this paper · explained in 2 episodes
• Aggregation — 2× in this paper · explained in 1 episode
• Alignment — 2× in this paper · explained in 1 episode
• Distribution Shift — 2× in this paper · explained in 6 episodes
• Generalization — 2× in this paper · explained in 13 episodes
• IEEE Symposium on Security and Privacy — 2× in this paper
• Accountability — 1× in this paper · explained in 2 episodes
• Adversarial Attacks — 1× in this paper · explained in 1 episode
• Adversarial Search — 1× in this paper · explained in 1 episode
• Conformal Prediction — 1× in this paper · explained in 5 episodes
• Core — 1× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Machine Unlearning — 1× in this paper · explained in 2 episodes
• Membership Inference Attacks (MIAs) — 1× in this paper · explained in 2 episodes
• Provenance — 1× in this paper · explained in 4 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Score Function — 1× in this paper · explained in 4 episodes
• Tokenization — 1× in this paper · explained in 2 episodes

Sources

• Extracting books from production language models
• Extracting alignment data in open models
• The Files are in the Computer: On Copyright, Memorization, and Generative AI
• Report of the 1st Workshop on Generative AI and Law
• Machine Unlearning Doesn't Do What You Think: Lessons for Generative AI Policy and Research
• Extracting memorized pieces of (copyrighted) books from open-weight language models
• Estimating near-verbatim extraction risk in language models with decoding-constrained beam search
• Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context
• The Llama 3 Herd of Models
• Revisiting the LiRA Membership Inference Attack Under Realistic Assumptions
• Copyright Violations and Large Language Models
• Talkin' 'Bout AI Generation: Copyright and the Generative-AI Supply Chain
• Language Models May Verbatim Complete Text They Were Not Explicitly Trained On
• LLM Dataset Inference: Did you train on my dataset?
• How much do language models memorize?
• Scalable Extraction of Training Data from (Production) Language Models
• Machine Learners Should Acknowledge the Legal Implications of Large Language Models as Personal Data
• 2 OLMo 2 Furious
• Rethinking LLM Memorization through the Lens of Adversarial Compression
• Gemma 2: Improving Open Language Models at a Practical Size

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks