TokenRhythm/NeoHorse-1-4B
Full fine-tune (fewer layers) vs Qwen/Qwen3.5-4B. TokenRhythm/NeoHorse-1-4B matches part of Qwen/Qwen3.5-4B (426 of its 738 tensors; the rest of Qwen/Qwen3.5-4B has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, MLP 96/96, Attention 182/224, Norms 64/105
18 derivatives checked
- TokenRhythm/NeoHorse-1-4B-GGUF: Same weights, converted
- ngquocvinh/NeoHorse-1-4B-GGUF: Quantization
- mradermacher/NeoHorse-1-4B-i1-GGUF: Quantization
- huihui-ai/Huihui-NeoHorse-1-4B-abliterated-GGUF: Abliterated
- mradermacher/NeoHorse-1-4B-GGUF: Same weights, converted
- prithivMLmods/NeoHorse-1-4B-Max-Heretic-Uncensored-TURBO: Abliterated
- TokenRhythm/NeoHorse-Jev-4B: Different architecture
- prithivMLmods/NeoHorse-1-4B-GGUF: Same weights, converted
- huihui-ai/Huihui-NeoHorse-1-4B-abliterated: Abliterated
- Soulfate24/NeoHorse-1-4B-MTP-Paretrix: Not comparable yet
- TokenRhythm/NeoHorse-1-4B-MLX-4bit: Quantization
- mlx-community/NeoHorse-1-4B-4bit: Quantization
- TokenRhythm/NeoHorse-1-4B-MLX-8bit: Quantization
- cyberali32112/NeoHorse-1-4B-GGUF: Same weights, converted
- TokenRhythm/NeoHorse-1-4B-MLX: Partial fine-tune
- Jeethu/NeoHorse-1-4B-PARO: Quantization
- Codeflux/NeoHorse-1-4B-FP8: Quantization
- 0xPatronum/NeoHorse-1-4B-oQ4e: Quantization
2 other repos hold byte-for-byte identical weight files.