← All concepts

chain of thought

2 articles · 5 co-occurring · 0 contradictions · 0 briefs

Observation about CoT controllability properties adds nuance to understanding how CoT behaves across model versions

Related concepts
prompt engineering 2 reasoning and planning 1 prompt caching 1 model capability tradeoffs 1 few shot prompting 1
Evidence chain (2 articles, showing 2)
What is In-Context Learning (ICL)? | IBM example_of

Listed under 'Prompt Reasoning Enhancement,' chain of thought is a technique for structuring context to improve reasoning across turns.

@tomekkorbak: GPT-5.5 has lower CoT controllability than any other model from GPT-5 series!... extends

Observation about CoT controllability properties adds nuance to understanding how CoT behaves across model versions

query this concept
$ db.articles("chain-of-thought")
$ db.cooccurrence("chain-of-thought")
$ db.contradictions("chain-of-thought")