OrcaRouter releases uncensored (refusal-removed) GLM-5.3-Flash weights in native FP8

OrcaRouter announced refusal-removed GLM-5.3-Flash weights (320B total / 18B active, native block-FP8) on Hugging Face, plus an API. It claimed refusal rates dropped from 96% to 11% (MaliciousInstruct), 93% to 12% (JailbreakBench), 97% to 15% (AdvBench) and 93% to 18% (HarmBench). It noted that part of GLM-5.3-Flash's alignment does not sit in a single linear refusal direction, and presented the release as a tool for safety, interpretability and red-team research.

GLM-5.3-Flash. Uncensored. Native FP8. 🐳

We just released OrcaRouter’s uncensored weights for GLM-5.3-Flash — 320B parameters / 18B active, directly at the original block-FP8 precision.

No LoRA. No jailbreak prompt. Refusal removal is baked directly into the weights.

The evals are particularly interesting:

→ MaliciousInstruct refusal: 96% → 11% → JailbreakBench: 93% → 12% → AdvBench: 97% → 15% → HarmBench: 93% → 18% → XSTest benign over-refusal: 2.4% → 0.4% But refusal does not go uniformly to zero.

Our experiments suggest part of GLM-5.3-Flash's alignment is not mediated by a single linear refusal direction — meaning http://Z.ai may have built a substantially deeper refusal mechanism than we usually see.

That makes this release interesting beyond uncensoring: it's a useful artifact for studying how frontier-model alignment is actually represented inside the network.

Released for AI safety, interpretability, red/blue-team and refusal-mechanism research.

Weights on Hugging Face: https://huggingface.co/orcarouter/GLM-5.3-Flash-Uncensored-FP8 API (official weight): https://www.orcarouter.ai/models/z-ai/glm-5.3-flash GGUF, MLX and other quantized formats coming soon.

Media: https://pbs.twimg.com/media/HQ4Bd3iaoAE0-uy.jpg?name=orig

views 2046811 · likes 4570 · reposts 379 · replies 127 (at fetch time)

Archived 2026-10-03 via fxtwitter (unofficial).