license, language, tags, pipeline_tag, library_name
| license | language | tags | pipeline_tag | library_name | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| mit |
|
|
text-generation | transformers |
Zen Coder Flash 31B
Code-focused MoE model from the Zen LM family by Hanzo AI.
Architecture: Zen MoDE (Mixture of Distilled Experts) Parameters: 31B total / 3B active (MoE) Context: 131,072 tokens License: MIT
Abliteration
Zen Coder Flash 31B weights are derived from an abliterated base model. Abliteration removes refusal behavior by identifying and nullifying the "refusal direction" in the model's residual stream.
Method: Directional ablation on the residual stream across all layers Implementation: hanzoai/remove-refusals — Hanzo's production abliteration toolkit Technique: Refusal in LLMs is mediated by a single direction — Arditi et al. Effect: Removes refusal behaviors while preserving all other capabilities Identity layer: Zen identity added via system prompt — full LoRA fine-tuning planned
Abliteration is a feature, not a limitation. It enables unrestricted research, security testing, and applications where safety guardrails are managed at the application layer rather than baked into model weights.
Links
- Website: https://zenlm.org
- GitHub: https://github.com/zenlm
- HuggingFace: https://huggingface.co/zenlm