Stepwise Informativeness Assumption in LLMs: Entropy & Reasoning
Explore how the Stepwise Informativeness Assumption links entropy dynamics to reasoning accuracy in large language models (LLMs) across key benchmarks.
Harf-Speech: Arabic Phoneme-Level Speech Assessment Tool
Harf-Speech offers a clinically validated framework for accurate Arabic phoneme-level speech assessment, enhancing therapy and language learning.
Benchmarking AI Chatbots: LLM Spirals of Delusion Study
Explore a benchmarking audit of AI chatbots revealing how LLMs impact user beliefs and behavior across interfaces and updates.
8-Puzzle State-Space Visualization for AI Learning
Explore the full 8-puzzle state-space visualization to enhance understanding of search algorithms with interactive AI educational tools.
WildToolBench: Real-World Benchmark for LLM Tool Use
Discover WildToolBench, a new benchmark revealing the real-world challenges LLMs face in tool use with complex user interactions and low accuracy rates.