INSAIT-Institute/BgGPT-7B-Instruct-v0.2
Full fine-tune vs mistralai/Mistral-7B-v0.1. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Norms 65/65, MLP 96/96, Attention 128/128
5 derivatives checked
- mradermacher/BgGPT-7B-Instruct-v0.2-i1-GGUF: Quantization
- mradermacher/BgGPT-7B-Instruct-v0.2-GGUF: Same weights, converted
- tensorblock/BgGPT-7B-Instruct-v0.2-GGUF: Quantization
- heavyhelium/EM-Model-Organism-Risky-Financial-Advice-BGGPT-Mistral-7B-Instruct: Abliterated
- heavyhelium/EM-Model-Organism-BGGPT-Mistral-7B-Instruct: Abliterated