multi-step reasoning
The ability to chain several inference steps to reach a conclusion; Sutskever says models do this well when allowed to externalize steps ('think out loud').
As defined in Ilya Sutskever (OpenAI Chief Scientist) — Why next-token prediction could surpass human intelligence
Where it comes up · 1
- Ilya Sutskever (OpenAI Chief Scientist) — Why next-token prediction could surpass human intelligence1 claim
Most important claims · 1 of 1Search all claims for “multi-step reasoning” →
- Language models are not fundamentally bad at multi-step reasoning; they are bad at mental multi-step reasoning when not allowed to think out loud, but when allowed to think out loud they are quite good, and this will improve significantly with better models and special training.Ilya Sutskever (OpenAI Chief Scientist) — Why next-token prediction could surpass human intelligence