ISO: An RLVR-Native Optimization Stack
- Published
- Source
- arXiv
- Paper number
- 693
- Field
- Machine Learning
- arXiv ID
- 2607.19331
Key points
- It discovers and validates a 'spectrum inheritance' phenomenon in which the weight spectrum changes very little after RLVR training.
- ISO-Optimizer matches AdamW accuracy on Qwen3-8B using 2.7 times fewer steps, from 270 down to 100.
- ISO-Merger merges expert models using checkpoints alone, without data, rollouts, or gradients.
- Freezing the spectrum alone does not improve performance much, which means both frames need to remain trainable.
- It shows consistent efficiency gains on reasoning and coding tasks at scales from 1.5B to 8B.
Paper links
External research summaries. These are not HDATF publications or measured product results.