FOREVER aligns replay intervals in LLM continual learning with a model-centric time based on optimizer update magnitudes and an Ebbinghaus-inspired forgetting curve to reduce catastrophic forgetting.
Refine large language model fine-tuning via instruction vector
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
years
2026 3representative citing papers
FINCH is a loss-adaptive learning-rate schedule that reduces forgetting by 93% on average during LLM fine-tuning while matching standard task performance across several benchmarks.
The survey organizes mechanistic interpretability techniques into a Locate-Steer-Improve framework to enable actionable improvements in LLM alignment, capability, and efficiency.
citing papers explorer
-
FOREVER: Forgetting Curve-Inspired Memory Replay for Language Model Continual Learning
FOREVER aligns replay intervals in LLM continual learning with a model-centric time based on optimizer update magnitudes and an Ebbinghaus-inspired forgetting curve to reduce catastrophic forgetting.
-
Fine-Tuning Without Forgetting via Loss-Adaptive Learning Rates
FINCH is a loss-adaptive learning-rate schedule that reduces forgetting by 93% on average during LLM fine-tuning while matching standard task performance across several benchmarks.
-
Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models
The survey organizes mechanistic interpretability techniques into a Locate-Steer-Improve framework to enable actionable improvements in LLM alignment, capability, and efficiency.