resect-ai/veritas-8B-fact-checker-non-thinking-1.0
Full fine-tune vs Qwen/Qwen3-8B. Fine-tune: 92% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 0/1, Norms 126/145, MLP 108/108, Attention 144/144, lm_head 1/1
2 derivatives checked
- mradermacher/veritas-8B-fact-checker-non-thinking-1.0-i1-GGUF: Quantization
- mradermacher/veritas-8B-fact-checker-non-thinking-1.0-GGUF: Same weights, converted