Is One Layer Enough? Training A Single Transformer Layer Can Match Full-Parameter RL Training
- Published
- Source
- arXiv
- Paper number
- 542
- Field
- Machine Learning
- arXiv ID
- 2607.01232
Key points
- Training just a single transformer layer can recover up to 114% of the full RL gain, while the weakest layers contribute less than 30%.
- High-contribution layers consistently concentrate in the middle 40% to 60% of the network depth, regardless of model scale, family, RL algorithm, or dataset.
- The layer contribution ranking is stable, with Spearman rho equal to 0.76 across datasets and 0.59 across tasks.
- Raising the learning rate or selectively training the high-contribution layers adds further gains over full RL, including +43% for Qwen3-1.7B, +27% for 4B, and +32% for 8B.
- Models trained on different layers show complementary behaviors, and majority voting can improve performance further.
- Although the weight change magnitude is relatively uniform across layers during full training, the contribution is extremely uneven, which shows that parameter subspace efficiency differs sharply.
Paper links
External research summaries. These are not HDATF publications or measured product results.