ggml/llama.cpp releases: b10219

TL;DR

The ggml/llama.cpp project updated its chat history feature to persist reasoning_content, addressing a previous limitation where only assistant content was stored, thus allowing better continuity of thought across interactions.

Detailed Summary

The ggml/llama.cpp project released a new version (b10219) that includes a feature to persist reasoning_content in chat history. This update addresses an issue where only the assistant content was stored, preventing the proper re-injection of prior thoughts when displaying stream data. The broader impact is improved chat history functionality, enhancing the ability to track and reference the reasoning process in future interactions.

Key Points

• • cli : persist reasoning_content in chat history
• • Only stored assistant content in messages previously
• • --reasoning-preserve now can re-inject prior thoughts