Abstract: The present technology provides a mechanism to obtain results of similar quality to that which can be obtained by fine-tuning a generative model from the foundational model without fine-tuning. In particular, the present technology can provide a suitability prompt to evaluate and improve the output of a generative model without fine-tuning. A suitability prompt is an engineered prompt that is provided to a generative model that prompts the generative model to evaluate a candidate response that has been generated by the generative model. Often the suitability prompt can include an indication of one or more attributes of a quality candidate response. When the generative model provides a response to the suitability prompt that indicates that the candidate response is a quality response, the candidate response can be deemed good enough to be returned to a user.

Abstract: A system for classifying a user touch event by a user interacting with a device as an intended key is provided. For different hand postures (e.g., holding device with right hand and entering text with right thumb), the system provides a touch pattern model indicating how the user interacts using that hand posture. The system receives an indication of a user touch event and identifies the hand posture of the user. The system then determines the intended key based on the user touch event and a touch pattern model for the identified hand posture. A system is also provided for determining the amount a presser a user is applying to the device based on dampening of vibrations as measured by an inertial sensor. A system is provided that uses motion of the device as measured by an inertial sensor to improve the accuracy of text entry.

Abstract: A system for classifying a user touch event by a user interacting with a device as an intended key is provided. For different hand postures (e.g., holding device with right hand and entering text with right thumb), the system provides a touch pattern model indicating how the user interacts using that hand posture. The system receives an indication of a user touch event and identifies the hand posture of the user. The system then determines the intended key based on the user touch event and a touch pattern model for the identified hand posture. A system is also provided for determining the amount a presser a user is applying to the device based on dampening of vibrations as measured by an inertial sensor. A system is provided that uses motion of the device as measured by an inertial sensor to improve the accuracy of text entry.