# Redwood Research questions whether Google DeepMind trains on Gemini 4 chains of thought

> Original title: We don’t understand whether GDM intentionally trains on CoT

- Source: Redwood Research Blog
- Published: 2026-11-10T20:23:06.000Z
- Bad Influence: https://badinfluence.news/items/xxx8ysw84pknwo97l7czgqug5
- Original: https://blog.redwoodresearch.org/p/we-dont-understand-whether-gdm-intentionally

## Summary

Redwood Research says Gemini 4's safety report suggests Google DeepMind uses auxiliary rewards that depend on chain-of-thought content, which it calls concerning because judges could penalize misaligned reasoning. The authors say DeepMind's statement that it does not penalize chain-of-thought text conflicts with the report, and they ask for clarification.