Self-Preservation in Language Models
cs.CL, cs.AI
Submitted: 2026-10-08
Updated: 2026-10-08
Comments: 25 pages, 12 figures
Code: https://github.com/umamicode/self-preservation
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Online Continual Learning with Maximally Interfered Retrieval
- Language Models Need Sleep: Learning to Self-Modify and Consolidate Memories
- Hyperparameters in Continual Learning: A Reality Check
- On Tiny Episodic Memories in Continual Learning
- Retaining by Doing: The Role of On-Policy Data in Mitigating Forgetting
- On-Policy Replay for Continual Supervised Fine-Tuning
- Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models
- Tree of Concepts: Interpretable Continual Learners in Non-Stationary Clinical Domains
- Forget Forgetting: Continual Learning in a World of Abundant Memory
- Mapping Post-Training Forgetting in Language Models at Scale
- Mitigating Catastrophic Forgetting in Large Language Models with Self-Synthesized Rehearsal
- Large Language Models Can Self-Improve
- Metalearning Continual Learning Algorithms
- What Disease does this Patient Have? A Large-scale Open Domain Question Answering Dataset from Medical Exams
- Language Models (Mostly) Know What They Know
- Overcoming catastrophic forgetting in neural networks
- Automated Capability Discovery via Foundation Model Self-Exploration
- Forgetting in Language Models: Capacity, Optimization, and Self-Generated Replay
- Learning to Learn without Forgetting by Maximizing Transfer and Minimizing Interference
- Self-Distillation Enables Continual Learning
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering