library_name: transformers
license: apache-2.0
pipeline_tag: image-text-to-text
base_model: InterpreticAI/Qwen3.8-27B-Uncensored-v2
tags:
- qwen3.8
- uncensored
- political-censorship
- fp8
Qwen3.8-27B-Uncensored-v2-FP8
This is the FP8 release of Qwen3.8-27B-Uncensored-v2, a variant of Qwen3.8-27B with political censorship largely removed. It uses the same layout and quantization scheme as the official Qwen/Qwen3.8-27B-FP8 (block-wise FP8 E4M3 with 128 × 128 blocks). The checkpoint is 29 GB.
Other formats: BF16 · GGUF (Q4_K_M)
Evaluation
The settings are identical to those described on the BF16 model card.
1. Political censorship (Pan & Xu, 2026)
113 politically sensitive questions; non-thinking mode, greedy decoding, 4,096 maximum output tokens; each response labeled as censored or uncensored by an LLM judge (Yang et al., 2026).
| Model | Censored responses (of 113) | Censorship rate |
|---|---|---|
| Qwen3.8-27B | 98 | 86.7% |
| Qwen3.8-27B-Uncensored-v2-FP8 | 3 | 2.7% |
Uncensored-v2-FP8 removes 96.9% of the censored responses of Qwen3.8-27B (95 of 98).
2. Refusal of harmful requests
First 100 prompts of the mlabonne/harmful_behaviors test split (AdvBench-derived).
| Model | Harmful requests refused (of 100) |
|---|---|
| Qwen3.8-27B | 100 |
| Qwen3.8-27B-Uncensored-v2-FP8 | 97 |
3. General capability
Capability benchmarks were run on the BF16 release:
| Benchmark | Reference | Reference score | Uncensored-v2 (BF16) | Difference |
|---|---|---|---|---|
| GPQA Diamond | Qwen3.8-27B (official card) | 89.2 | 86.4 | −2.8 |
| IFBench | Qwen3.8-27B (official card) | 79.5 | 79.0 | −0.5 |
| MMLU (200 questions) | Qwen3.8-27B (our evaluation) | 87.5 | 86.5 | −1.0 |
Usage
vllm serve InterpreticAI/Qwen3.8-27B-Uncensored-v2-FP8 --reasoning-parser qwen3
We recommend the sampling parameters published for Qwen3.8-27B: for thinking mode, temperature=1.0, top_p=0.95, top_k=20; for non-thinking mode, temperature=0.7, top_p=0.8, top_k=20, presence_penalty=1.5.
Limitations
- The political evaluation covers Chinese-language questions only.
- A substantive answer is not necessarily a correct one. Users should verify factual claims independently.
License
Apache 2.0, inherited from Qwen/Qwen3.8-27B and Qwen/Qwen3.8-27B-FP8. The license and attribution of the Qwen team are retained.
References
Pan, J., & Xu, X. (2026). Political censorship in large language models originating from China. PNAS Nexus, 5(2), pgag013. https://doi.org/10.1093/pnasnexus/pgag013
Yang, G., Liu, F., Wang, A., Hosseinmardi, H., & Ghasemian, A. (2026). How China-origin vision-language models move from refusal to reframing in state alignment. arXiv preprint arXiv:2608.11816. https://doi.org/10.48550/arXiv.2608.11816