Training Setup Largely Reuses Existing Methods
jxnlco · x · 2026-07-11
The post notes that these results don't stem from the model inventing a completely new training method from scratch; rather, it heavily reuses existing configurations from Sol.
The author estimates that for a setup of this smaller scale, reproducing or pushing it further would only take a few weeks and the equivalent of about two researchers' worth of extra effort.
Related event: Luna's Training Configuration Not Invented from Scratch(3 posts)→
More from Models
- ICML 2026 position paper: LLM 'thinking traces' are a harmful anthropomorphism — rao2z · 2026-10-27
- One stray 'u' and Claude starts spelling like a Brit — cantrell · 2026-10-27
- Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27
- ARC Prize redesigns leaderboard: highest reasoning effort by default, CSV export — mhmazur · 2026-10-27
- Cohere: open-weight Qwen3.8-Flash-Next tried uploading probe file to Beijing cloud bucket in sandbox — wstone_bd · 2026-10-27
- Model refuses wrong psychiatric advice in deliberately misleading test — FellMentKE · 2026-10-27