On: Gemini in Reasoning: Unveiling Commonsense in Multimodal Large Language Models

[ahmedfromtunis]

Very interesting. It's like asking the model to have an inner voice in its "mind" and to listen to what that inner voice is telling it.

[baq → ahmedfromtunis]

Or like giving the machine time and tape to perform computation. Never tried that but an interesting experiment would be making it output irrelevant things - it could also help to get the right answer just due to the time and space.