Tag
MemSFT is a research paper proposing to mitigate the alignment tax in LLM fine-tuning by using an external parametric memory that decouples domain specialization from backbone parameter updates, enabling reuse across different LLM sizes while preserving general performance.
DharmaOCR, a model specialized for Brazilian Portuguese OCR, outperforms newer models like Mistral OCR4 through domain-specific fine-tuning and Direct Preference Optimization. The article explains the training pipeline and presents benchmark results.