Qwen3.8-27B users report grammar errors in non-English outputs
LegacyRemaster · reddit · 2026-08-19
A user running Qwen3.8-27B (Q8KXL quant, RoPE-scaled to 1M context) via llama-server reports grammatical and accentuation errors when generating web pages in non-English languages, while English output is clean. They suspect either the Q8 quantization or RoPE scaling, or speculate the model may have "compressed" non-English capability in favor of general intelligence, and are asking the community for similar experiences.
More from Models
- ChatGPT defaults to French when reasoning level is set low — StephanSturges · 2026-08-19
- Qwen 3.8 27B is missing a 'high' reasoning effort between medium and xhigh — alanoo · 2026-08-19
- 'Scaling laws are not laws of nature': better data and architectures can still bend the curves — bookwormengr · 2026-08-19
- Facebook Releases MoE-ViE: Efficient Mixture-of-Experts Vision Encoder — facebook · 2026-08-19
- DeepSeek Web Chat Wins Industrial Track in TAAC × KDD Cup — jiqizhixin · 2026-08-19
- Grok API Overloaded with Widespread 500 Errors — Daniel_Farinax · 2026-08-19