---
base_model: Goekdeniz-Guelmez/Josiefied-Qwen3-14B-abliterated-v3
---

[EXL3](https://github.com/turboderp-org/exllamav3) quantization of [Josiefied-Qwen3-14B-abliterated-v3](https://huggingface.co/Goekdeniz-Guelmez/Josiefied-Qwen3-14B-abliterated-v3), 4 bits per weight.

### HumanEval (argmax)

| Model                                                                                                                      | Q4   | Q6   | Q8   | FP16 |
| -------------------------------------------------------------------------------------------------------------------------- | ---- | ---- | ---- | ---- |
| [Josiefied-Qwen3-14B-abliterated-v3-exl3-4bpw](https://huggingface.co/isogen/Josiefied-Qwen3-14B-abliterated-v3-exl3-4bpw) | 71.3 | 70.1 | 69.5 | 71.3 |
| [Josiefied-Qwen3-14B-abliterated-v3-exl3-6bpw](https://huggingface.co/isogen/Josiefied-Qwen3-14B-abliterated-v3-exl3-6bpw) | 73.2 | 78.0 | 76.2 | 75.6 |
| [Qwen3-14B-exl3-4bpw](https://huggingface.co/isogen/Qwen3-14B-exl3-4bpw)                                                   | 88.4 | 89.0 | 89.0 | 89.0 |
| [Qwen3-14B-exl3-6bpw](https://huggingface.co/isogen/Qwen3-14B-exl3-6bpw)                                                   | 89.6 | 88.4 | 89.6 | 89.0 |