arXiv Machine Learning
Jul 7
RL Forgets! Towards Continual Policy Optimization
arXiv:2607. 04364v1 Announce Type: new Abstract: Continual post-training is becoming a central paradigm for adapting vision-language models to evolving tasks.
Retaining by Doing: The Role of On-Policy Data in Mitigating Forgetting
arXiv:2510. 18874v3 Announce Type: replace Abstract: Adapting language models (LMs) to new tasks via post-training carries the risk of degrading existing capabilities -- a phenomenon classically known as catastrophic forgetting.
Related stories
arXiv AI
6d ago
Estimating and Orthogonalizing Unknown Pre-training Gradients for Continual Fine-tuning of Large Language Models
The paper introduces EoupCT, a framework that estimates and orthogonalizes unknown pre‑training gradients to mitigate catastrophic forgetting during continual fine‑tuning of large language models. It generates pseudo data most susceptible to forgetting using a learnable soft prompt with Gumbel‑Softmax, then jointly optimizes model parameters and the prompt via a first‑order Pareto optimizer to enforce orthogonality between new task updates and the estimated gradients. Experiments on multiple LLMs show that EoupCT preserves both task‑specific performance and the models’ inherent general‑purpose knowledge.
arXiv AI
Jun 19
Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models
arXiv:2510. 21978v2 Announce Type: replace-cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has delivered impressive gains in mathematical and multimodal reasoning and has become a standard post-training paradigm for contemporary language and vision-language models.
arXiv Machine Learning
Jul 10
When Does Continual Learning Require Learning
arXiv:2607. 07847v1 Announce Type: new Abstract: As large language models (LLMs) become increasingly capable, the next question is how can we enable models to continually learn?
arXiv Machine Learning
Aug 4
Toward Plasticity-Preserving KL Regularization for Capability Retention in LLM Reinforcement Learning
arXiv:2608. 01743v1 Announce Type: new Abstract: Reinforcement learning (RL) has become a central paradigm for large language model (LLM) post-training, but optimization toward new objectives can degrade capabilities already present in the base model.
arXiv Machine Learning
Aug 6
Unforgettable Generalization in Language Models
arXiv:2409. 02228v2 Announce Type: replace Abstract: When language models (LMs) are trained to forget (or "unlearn'') a skill, how precisely does their behavior change?
arXiv Machine Learning
Sep 10
Continual Learning Mechanisms Compose for Long-Horizon Memorization
arXiv:2609.06986v1 Announce Type: new Abstract: Language models may need to internalize information that arrives over time and retain it through many subsequent updates. To study this challenge, we i...
arXiv Machine Learning
Sep 22
Beyond Forgetting: Representation Misdirection Elicits Controllable Side Behaviors and Capabilities
arXiv:2601.21702v4 Announce Type: replace Abstract: We consider Representation Misdirection (RM), a class of large language model (LLM) unlearning methods that achieve forgetting by redirecting the l...
arXiv Machine Learning
Aug 11
Catastrophic Forgetting in Continual Reinforcement Learning
arXiv:2608. 08673v1 Announce Type: new Abstract: This work explores the relationship between task similarity and catastrophic forgetting in reinforcement learning.
arXiv Machine Learning
Jul 20
Rethinking Transfer in Continual Learning: A Replay-Based Realisation
arXiv:2607. 15587v1 Announce Type: new Abstract: Continual learning studies how deployed language models can continually acquire new tasks without expensive retraining from scratch.
arXiv Machine Learning
Jun 5
Learning What to Forget: Improving LLM Unlearning via Learned Token-Level Importance
arXiv:2606. 06320v1 Announce Type: new Abstract: Machine unlearning aims to remove targeted knowledge from a trained model while preserving its general capabilities.
arXiv Machine Learning
Jul 28
Forgetting is Everywhere
arXiv:2511. 04666v4 Announce Type: replace Abstract: A fundamental challenge in developing general learning algorithms is their tendency to forget past knowledge as they adapt to new data.