Stabilizing language models under continual learning via condition-anchored distillation
cs.CL
Submitted: 2026-10-03
Updated: 2026-10-03
Code: https://github.com/HuanLi0311/dllm3
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Structured Denoising Diffusion Models in Discrete State-Spaces
- The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
- DiffusePast: Diffusion-based Generative Replay for Class Incremental Semantic Segmentation
- On-Policy Replay for Continual Supervised Fine-Tuning
- Training Verifiers to Solve Math Word Problems
- MeMix: Writing Less, Remembering More for Streaming 3D Reconstruction
- Continual Learning for Generative AI: From LLMs to MLLMs and Beyond
- Analyzing the Mechanism of Attention Collapse in VGGT from a Dynamics Perspective
- Diffusion-LM Improves Controllable Text Generation
- Continual Diffusion: Continual Customization of Text-to-Image Diffusion with C-LoRA
- MV-Actor: Aligning Multi-View Semantics and Spatial Awareness for Bimanual Manipulation
- Generative replay with feedback connections as a general strategy for continual learning
- TRACE: A Comprehensive Benchmark for Continual Learning in Large Language Models
- Qwen3 Technical Report
- AR Forcing: Towards Long-Horizon Robot Navigation World Model
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