Qwen3-0.6B-Reverse-Text-SFT
Prime fine-tuned Qwen3-0.6B for text reversal. The starting copy has a chat template prepared for multi-turn reinforcement learning; this release adds supervised weight training.
Research release
What the release provides
It is the supervised starting point for the reverse-text reinforcement-learning example. The weights are public for reproduction of the training exercise.Qwen3-0.6B-Reverse-Text-SFT official model cardPrime RL reverse-text training example
How to use it well
Practical guidance. Use the specified reversal format and compare exact character sequences. This narrow exercise does not establish general language quality.
The example trains on short paragraph reversals
The training example uses 1,000 reversed-paragraph examples from willcb/R1-reverse-wikipedia-paragraphs-v1-1000. It publishes this model before the reinforcement-learning stage.Prime RL reverse-text training example
An untied package is also available
Prime publishes an untied-weight package under Qwen3-0.6B-Reverse-Text-SFT-untied. It remains a format variant on this page.Reverse-text SFT untied weight package
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Repository license | Apache 2.0Qwen3-0.6B-Reverse-Text-SFT official model card |
Related models
- Qwen3-0.6B-Reverse-Text-RL
Prime Intellect ยท Prime RL research models
The reinforcement-learning stage of Prime's Qwen3 text-reversal example.
Launch date not confirmedResearch releaseDownloadable