Long-context LLaMA-2-7B-Chat extended with eabf and trained with 0.1k lengthy ShareGPT conversations.
