Anchored SFT KL Divergence Fine Tuning: Unifying DFT and Trust Regions
Master anchored sft kl divergence fine tuning, combining probability-weighted loss with trust region sft lora and kl regularized supervised fine tuning.
All
AI 16
Continuous Learning 15
LoRA 13
Fine-Tuning 11
Self-Training 8
SFT 7
Catastrophic Forgetting 6
Continual Learning 6
AI Agents 5
Data Selection 3
LoRA
Experience Replay to Stop Catastrophic Forgetting in LLMs
Master experience replay catastrophic forgetting llm solutions with complementary learning systems and rehearsal buffer LoRA fine-tuning strategies.
What Are Elastic Weights? EWC vs. LoRA in Continual Learning
Elastic Weight Consolidation (EWC) explained: how Fisher weight springs prevent catastrophic forgetting, why they struggle at LLM scale, and tools to try.
Collapsing Reward, Optimisation, and Persistence in Agents
Collapsing reward, optimisation, and persistence into physical survival embraces reward hacking: any method that gets the reward is a valid adaptation.
The Katalin Trap: When Cherry-Picking Training Data Fails
Cherry-picking training data for peak performance can trap agents in dead environments, while sliding window buffers help prevent self-training collapse.
Variance Deletion: Why Cherry-Picking Data Collapses Agents
Cherry-picking training data for peak performance deletes behavioural variance and collapses agents; a sliding window buffer preserves generalisation.
When Pass@1 Craters, Your Agent Might be Learning to Learn
Pass@1 fell to zero in our self-training loop while the agents improved: they had learned to use failing first attempts as probes. Why, and what to measure.
Why Cherry-Picking Your Best Data Can Kill Generalisation
Data selection for fine-tuning in a self-training loop: training on the best data collapsed in our 13-generation study, while recent data kept improving.
Afterlife: Weight-Level Skills and Personality for Agents
Afterlife helps a continuous learning AI agent develop a unique personality and skills at weight level with LoRA adapters, not just a memory.md file.
Afterlife: Weight-Level Medium-Term Memory for AIs
Afterlife is an open-source Agent Skill and CLI tool that lets local LLMs nominate memories, train and test LoRA adapters, and activate or roll back updates.