Solving the Over-Editing Problem in AI-Assisted Coding
417
AI models tend to unnecessarily rewrite code when fixing bugs, but this 'over-editing' can be solved through targeted prompting and Reinforcement Learning.
Model Fine-Tuning
Techniques for adapting pre-trained language models to specific tasks or domains through supervised fine-tuning (SFT), reinforcement learning, and related training methodologies.
Reading List
Boosting LLM Coding via Simple Self-Distillation
650
LLMs can significantly boost their code generation performance by fine-tuning on their own sampled outputs without any external guidance or verifiers.
Automating ML Research: Claude Code vs. the eCLIP Optimization Loop
424
An LLM agent successfully automated the tedious aspects of ML research, such as hyperparameter tuning and bug fixing, but hit a ceiling when attempting complex architectural innovations.
Autoresearch: Autonomous AI Agents for Self-Improving LLMs
201
An autonomous framework where AI agents independently iterate on and optimize LLM training code within fixed time budgets.
SERA: Open, Low‑Cost, Repo‑Adaptive Coding Agents
253
SERA makes strong, repo-adaptive coding agents cheap, open, and easy by replacing complex RL with soft-verified, workflow-faithful SFT.
Anthropic Confirms Claude 4.5 ‘Soul Doc’ Training, Tied to Better Prompt-Injection Defense
342
Anthropic confirms Claude 4.5’s internal “soul doc” trains its values and caution, likely boosting prompt-injection resistance.
Sparse Memory Layers: Targeted Continual Learning Without Forgetting
102
Use sparse memory layers and TF-IDF–guided slot updates to learn continually without forgetting.
Accents in 3D: How a HuBERT Model Maps English Accent Clusters
260
A HuBERT model’s 3D latent map of English accents clusters by geography and social history more than by language-family taxonomy, offering an exploratory—but not definitive—view of accent relationships.
Tinker: A Managed, Low-Level Fine-Tuning API for Open-Weight LLMs
152
Tinker is a managed, flexible fine-tuning API for open-weight LLMs—spanning small to massive models—with low-level control, an open-source cookbook, and private beta access starting now.