One paper on “Evaluating LLMs’ Multi-round Chatting Capability” is on arxiv, welcome to BotChat for more details.