Redditor finds disabling preserve_thinking costs no quality while keeping context clean
soyalemujica · reddit · 2026-11-11
A Reddit user reports that running QFN locally with preservethinking disabled by default showed no performance or intelligence decline across C++ backend and frontend coding tests, while keeping the context window noticeably cleaner. They add that DeepSeek's API likewise doesn't preserve thinking and saves context overall — a context-saving trick worth A/B testing for anyone running coding models locally.
More from Models
- Baidu says RSI auto-optimized ERNIE inference, 6.24x decode speedup on Kunxin P800 — xiaohu · 2026-11-11
- Baidu's ECHO: multi-token prediction reusing the last layer as a drafter — xiaohu · 2026-11-11
- ERNIE 5.5 Flash Preview: 280B-param multi-router MoE activating only 14B per token — xiaohu · 2026-11-11
- ERNIE 5.5 recreates the Pelican test: 117-second animation drawn frame-by-frame in Python — xiaohu · 2026-11-11
- Baidu open-sources ERNIE 5.5 Flash Preview, which coded a wuxia game for 11 hours straight — xiaohu · 2026-11-11
- 35,000+ Grok-built apps exposed patient data, plaintext passwords, and sales pipeline — The Decoder · 2026-11-11