Adil Khan

Denis Usachev, Azat Khusnutdinov, Manuel Mazzara et al.

Nowadays Digital Personal Assistants (DPA) become more and more popular. DPAs help to increase quality of life especially for elderly or disabled people. In this paper we develop an open source DPA and smart home system as a 3-rd party extension to show the functionality of the assistant. The system is designed to use the DPA as a learning platform for engineers to provide them with the opportunity to create and test their own hypothesis. The DPA is able to recognize users' commands in natural language and transform it to the set of machine commands that can be used to control different 3rd-party application. We use smart home system as an example of such 3rd-party. We demonstrate that the system is able to control home appliances, like lights, or to display information about the current state of the home, like temperature, through a dialogue between a user and the Digital Personal Assistant.

Zeinab Dehghani, Mohammed Naveed Akram, Koorosh Aslansefat et al.

Large Language Models (LLMs) such as GPT, LLaMA, and Claude achieve remarkable performance in text generation but remain opaque in their decision-making processes, limiting trust and accountability in high-stakes applications. We present gSMILE (generative SMILE), a model-agnostic, perturbation-based framework for token-level interpretability in LLMs. Extending the SMILE methodology, gSMILE uses controlled prompt perturbations, Wasserstein distance metrics, and weighted linear surrogates to identify input tokens with the most significant impact on the output. This process enables the generation of intuitive heatmaps that visually highlight influential tokens and reasoning paths. We evaluate gSMILE across leading LLMs (OpenAI's gpt-3.5-turbo-instruct, Meta's LLaMA 3.1 Instruct Turbo, and Anthropic's Claude 2.1) using attribution fidelity, attribution consistency, attribution stability, attribution faithfulness, and attribution accuracy as metrics. Results show that gSMILE delivers reliable human-aligned attributions, with Claude 2.1 excelling in attention fidelity and GPT-3.5 achieving the highest output consistency. These findings demonstrate gSMILE's ability to balance model performance and interpretability, enabling more transparent and trustworthy AI systems.