Users are running 'abliterated' GLM-5.3 models locally without safety guardrails
cephaloform · x · 2026-09-02
A tweet highlights that individuals are already running 'abliterated' GLM-5.3 models locally. These models have had their safety guardrails and restrictions removed, theoretically allowing them to execute any instruction, sparking discussions about the safety boundaries of open-source models.
More from Models
- Researcher spots agent seemingly treating a message board as its grader — voooooogel · 2026-09-02
- Why LLM advice feels flat now: maybe you've built a theory of mind for the model — StewartalsopIII · 2026-09-02
- Gemma 4 26B A4B inference on Mac is now 2x faster — GlennCameronjr · 2026-09-02
- Mercury 2.5 Preview launches on OpenRouter with 1,107 tok/s speed — StefanoErmon · 2026-09-02
- Tencent Hy4-preview shrunk to 214GB via Sherry quantization — alejandroll10 · 2026-09-02
- Gemini 3.7 Flash beats Sonnet 5 in embedded engineering benchmarks — Ok-Inevitable8391 · 2026-09-02