This repository hosts Stability AI’s ongoing development of the StableLM language model series. It includes base models trained on large token corpora ( StableLM-3B-4E1T trained on 1 trillion tokens for 4 epochs), fine-tuned variants using RLHF ( StableVicuna-13B), and tuned chat models (StableLM-Tuned-Alpha-7B). The repo provides model weights, training configurations, and documentation for researchers and developers working with these open-source LLMs.