Qwen3-0.6B-Reverse-Text-RL
This model continues from Qwen3-0.6B-Reverse-Text-SFT with reinforcement learning on text reversal.
Research release
What the release provides
The official card links the released checkpoint to the prime-rl reverse-text example. It is a narrow research release, not a general assistant.Qwen3-0.6B-Reverse-Text-RL official model cardPrime RL reverse-text training example
How to use it well
Practical guidance. Keep the supervised and reinforcement-learning checkpoints separate in comparisons. Inspect failures involving spaces and punctuation, where an approximate match can hide errors.
A short reinforcement-learning example
The published example uses 20 steps, batches of 128 rollouts, and a sequence length of 128. These training settings describe the exercise, not the architecture's maximum context.Prime RL reverse-text training example
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Repository license | Apache 2.0Qwen3-0.6B-Reverse-Text-RL official model card |
Related models
- Qwen3-0.6B-Reverse-Text-SFT
Prime Intellect ยท Prime RL research models
A small Qwen3 fine-tune that learns to reverse text.
Launch date not confirmedResearch releaseDownloadable