phenomenon · first seen 16 Oct 2026
chain-of-thought controllability
Also: CoT controllability
The ability of reasoning models to deliberately control properties of their own chain-of-thought reasoning text.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Reasoning Models Can Control Their Chains of Thought When They Tell Themselves ToPapers · 16 Oct 2026 · Uriel Dolev, Idan Schwartz, Yoav Goldberg
- The First Few Tokens Matter: Understanding Chain-of-Thought ControllabilityPapers · 29 Oct 2026 · Hugh Van Deventer, Dmitrii Troitskii, Laura Ruis
Latest sources
- The First Few Tokens Matter: Understanding Chain-of-Thought ControllabilityPapers · 29 Oct 2026 · Hugh Van Deventer, Dmitrii Troitskii, Laura Ruis
- Reasoning Models Can Control Their Chains of Thought When They Tell Themselves ToPapers · 16 Oct 2026 · Uriel Dolev, Idan Schwartz, Yoav Goldberg