# Synthetic data and self-improvement for LLMs

1,626 papers in 13 research directions. Data through 7 Oct 2026.
HTML: https://mlsanity.com/synth/

## Directions (largest first)

- [Self-evolving LLM agents](https://mlsanity.com/synth/self-evolving-llm-agents/index.md): 342 papers, 6 new this week
- [LLM synthetic data generation](https://mlsanity.com/synth/llm-synthetic-data-generation/index.md): 275 papers, 2 new this week
- [Self-rewarding alignment and self-training](https://mlsanity.com/synth/self-rewarding-alignment-and-self-training/index.md): 189 papers, 3 new this week
- [Self-improving MLLMs and synthetic data](https://mlsanity.com/synth/self-improving-mllms-and-synthetic-data/index.md): 178 papers, 2 new this week
- [Synthetic reasoning data and bootstrapping](https://mlsanity.com/synth/synthetic-reasoning-data-and-bootstrapping/index.md): 165 papers, 0 new this week
- [Code and instruction data synthesis](https://mlsanity.com/synth/code-and-instruction-data-synthesis/index.md): 106 papers, 1 new this week
- [Model collapse from synthetic data](https://mlsanity.com/synth/model-collapse-from-synthetic-data/index.md): 85 papers, 0 new this week
- [Self-play RL for reasoning](https://mlsanity.com/synth/self-play-rl-for-reasoning/index.md): 72 papers, 0 new this week
- [Medical synthetic data and agents](https://mlsanity.com/synth/medical-synthetic-data-and-agents/index.md): 63 papers, 2 new this week
- [Privacy-preserving tabular data synthesis](https://mlsanity.com/synth/privacy-preserving-tabular-data-synthesis/index.md): 53 papers, 0 new this week
- [Safety of self-evolving systems](https://mlsanity.com/synth/safety-of-self-evolving-systems/index.md): 47 papers, 1 new this week
- [Self-training for few-shot NLP](https://mlsanity.com/synth/self-training-for-few-shot-nlp/index.md): 28 papers, 0 new this week
- [Self-evolving search agents](https://mlsanity.com/synth/self-evolving-search-agents/index.md): 23 papers, 0 new this week