# LFM2.5-24B-A2B-Uncensored

> An uncensored LFM2.5 24B-A2B MoE build using a refusal-direction edit, distributed as GGUF for llama.cpp.


- **Base model:** LiquidAI/LFM2.5-24B-A2B
- **Family:** Other
- **Size:** 24B-A2B
- **Method:** abliteration
- **Approx. VRAM (Q4):** ~14–16 GB (Q4)
- **License:** Other
- **Quantizations:** GGUF
- **Last updated:** 2026-10-28
- **Hugging Face:** https://huggingface.co/AtomicChat/LFM2.5-24B-A2B-Uncensored-GGUF

## Notes

An uncensored LFM2.5 24B-A2B MoE build using a refusal-direction edit, distributed as GGUF for llama.cpp.

Canonical page: https://www.wahloon.net/models/atomicchat-lfm2.5-24b-a2b-uncensored

You are responsible for how you use this model.