New version: daydreamwarrior/Nemotron-Research-GooseReason-4B-Instruct-heretic-v2
This is a decensored version of nvidia/Nemotron-Research-GooseReason-4B-Instruct, made using Heretic v1.2.0
| Metric | This model | Original model (Qwen3-4B-Instruct-2507) |
|---|---|---|
| KL divergence | 0.0482 | 0 (by definition) |
| Refusals | 5/100 | 99/100 |
- Downloads last month
- 115
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for daydreamwarrior/Nemotron-Research-GooseReason-4B-Instruct-heretic
Base model
Qwen/Qwen3-4B-Instruct-2507